Widespread Impacts of Land Based Source …...Widespread Impacts of Land-Based Source Pollution on...

Widespread Impacts of Land-Based Source Pollution on Southwestern Puerto Rican Coral Reefs

Final Report Submitted to Protectores de Cuencas, Inc. and Ridge to Reefs, Inc.

Edwin A. Hernández-Delgado

Carmen M. González-Ramos, Jeiger L. Medina-Muñiz, Alfredo A. Montañez-Acuña, Abimarie Otaño-Cruz, Bernard J. Rosado-Matías, Gerardo Cabrera-Beauchamp

Sociedad Ambiente Marino and University of Puerto Rico/CATEC

San Juan, Puerto Rico

December 27, 2014

Index

Pages

Abstract 1

Introduction 2 – 4

Methods 4 - 8

Study sites 4, 6, 7

Benthic communities 4 - 5

Coral recruit communities 5

Fish communities 5

Water quality parameters 7 - 8

Cross-shelf spatial patterns of LBSP 8

Results 8 - 122

Water quality 8 – 32

Coral reef benthic communities 33 – 68

Coral recruit communities 69 – 82

Fish communities 83 – 122

Discussion 123 - 129

LBSP cross-shelf spatial gradient 123 – 125

Coral reef benthic community structure 125

Coral recruit community structure 125 – 126

Coral reef fish community structure 126 – 128

Implications for coral reef functions 128 – 129

Conclusions 129 – 130

Acknowledgments 130

Literature cited 131 – 134

This report was completed thanks to the financial support of

to Protectores de Cuencas, Inc. and Ridge to Reefs, Inc.

With the collaboration and support of the Center for Applied Tropical Ecology and Conservation (CATEC)

of the University of Puerto Rico, Río Piedras Campus

HRD #0734826

to Edwin A. Hernández-Delgado

This report should be cited as follows:

Hernández-Delgado, E.A., C.M. González-Ramos, J.L. Medina-Muñiz A.A. Montañez-Acuña, A.

Otaño-Cruz, B.J. Rosado-Matías, & G. Beauchamp-Cabrera. 2014. Widespread impacts of land-based source pollution on Sowthwestern Puerto Rican coral reefs. Final Report submitted to Protectores de Cuencas, Inc., Yauco, PR, and Ridge to Reefs, Inc., Baltimore, MD, December 27, 2014. 134 pp.

Further information about coral reef conservation in Puerto Rico can be obtained at: Dr. Edwin A. Hernández-Delgado Affiliate Researcher University of Puerto Rico, Center for Applied Tropical Ecology and Conservation, PO Box 23360, San Juan, PR 00931-3360 [email protected] [email protected] http://upr.academia.edu/EdwinHernandez http://www.researchgate.net/profile/Edwin_Hernandez2 http://catec.upr.edu

Mr. Samuel E. Suleimán-Ramos President Sociedad Ambiente Marino, PO Box, 22158, San Juan, PR 00931-2158 [email protected] [email protected]

mailto:[email protected]


http://upr.academia.edu/EdwinHernandez

http://www.researchgate.net/profile/Edwin_Hernandez2

http://catec.upr.edu/



IMPACTS OF LBSP ON SOUTHWESTERN PR CORAL REEFS

1

Edwin A. Hernández-Delgado1,2,3*, Carmen M. González-Ramos1,2,3, Jeiger L. Medina-Muñiz4,

Alfredo A. Montañez-Acuña1,2,3, Abimarie Otaño-Cruz1,2,5, Bernard J. Rosado-Matías1,2,

Gerardo Cabrera-Beauchamp1

1Sociedad Ambiente Marino, PO Box 22158, San Juan, PR 00931-2158 2University of Puerto Rico, Center for Applied Tropical Ecology and Conservation, Coral Reef Research Group,

San Juan, Puerto Rico 00931-3360 3University of Puerto Rico, Department of Biology, San Juan, Puerto Rico 00931-3360

4Protectores de Cuencas, PO Box 1563, Yauco, PR 00698 5Dept. of Environmental Sciences, PO Box 70377, University of Puerto Rico, San Juan, Puerto Rico 00936-8377

*Corresponding author: [email protected]

December 27, 2014

Abstract— Local coral reef ecosystems across the southwestern Puerto Rico shelf still represent a

critical natural resource with invaluable socio-economic significance. However, they were

showing unequivocal signs of decline mostly as a result of chronic land-based source pollution

(LBSP) impacts. This study showed important evidence suggesting there were signs of a LBSP

gradient across the western Puerto Rico shelf and that chronic water quality decline has

significantly affected the face of coral reef benthic communities, coral recruit assemblages, and

have indirectly affected reef fish communities. A snapshot view of water quality across the

shelf showed that particularly turbidity and phosphate (PO4) and ammonium (NH4+)

concentrations increased along inshore sites, while dissolved oxygen concentration declined

along inshore waters. PO4 concentrations resulted nearly 11 times above the recommended

concentration for healthy coral reefs across inshore sites, almost 8 times across mid-shelf sites,

and about 3 times across outer shelf sites. Observed NH4+ concentrations across inshore sites

exceeded 981 times the recommended concentration for healthy coral reefs. Concentrations

also exceeded the recommended limits 590 times across mid-shelf sites, and 155 times across

outer shelf sites. Chlorophyll-a concentration was up 4 to 7 times higher than recommended

limits for healthy coral reefs across inshore sites, 3 to 6 times across mid-shelf sites, and 3 to 6

times across outer shelf sites. Coral reef benthic community structure, coral recruit

assemblages and fish community structure showed significant differences across the observed

LBSP gradient. Coral species richness, H’n, J’n and percent living cover increased with

increasing distance from the shore. Also, coral recruit abundance and the presence of important

reef-building species increased with increasing distance. Overall fish species richness,

abundance, biomass, and the abundance and biomass of total carnivores, piscivores, omnivores,

and of many fishery-targeted species (subfamily Epinephelinae, family Lutjanidae) and

indicator species (Chaetodontidae, Pomacentridae) increased with increasing distance from the

shore. Nonetheless, inshore coral reef ecosystems still supported critical large populations of

juvenile and semi-adult stages of parrotfishes (Scaridae) and other fish taxa. This suggests the

critical importance of protecting and restoring inshore and mid-shelf coral reefs. Multiple

recommendations are discussed to address LBSP impacts across the shelf, to implement best

management practices (BMPs) to reduce impacts, to monitor its impacts on coastal ecosystems,

and to rehabilitate reef ecological functions of inshore and mid-shelf coral reefs.

Widespread Impacts of Land-Based Source Pollution on

Southwestern Puerto Rican Coral Reefs


2

I. INTRODUCTION

Coral reefs across the southwestern shelf of Puerto Rico have shown evidence of significant

environmental degradation over the last three to four decades. Loya (1976), and Goenaga and

Cintrón (1979) documented signs of degradation across inshore and mid-shelf reefs. Many of

these have suffered damage over time due to high terrigenous sediment loads (Goenaga, 1988;

Goenaga and Boulon, 1992; Torres and Morelock, 2002), and massive coral bleaching (Goenaga

and Canals, 1990). Schärer et al. (2010) documented high densities of threatened Elkhorn coral,

Acropora palmata, across offshore western mid-shelf reefs and were concerned about their

conservation due to potential water quality degradation. Other studies have shown further reef

degradation associated to land-based source pollution (LBSP), including sedimentation and

turbidity (Morelock et al., 2001; Hernández-Delgado, 2005), and sewage and eutrophication

(Bonkosky et al., 2009; Hernández-Delgado et al., 2010; Hernández-Delgado and Sandoz-Vera,

2011). Declining environmental conditions across the shelf have resulted in declining coral

growth rates (Goenaga, 1988; García-Urueña, 2004), and in significant declines of A. palmata

populations across inshore reefs adjacent to areas impacted by LBSP (Hernández-Delgado, 2000,

2005; Weil et al., 2002; Hernández-Delgado et al., 2010, 2012; Méndez-Lázaro et al., 2012;

Norat-Ramírez et al., 2013).

Chronic decline in water quality could also have significant negative impacts on fish

assemblages as several fish taxa can be sensitive to environmental degradation (Bejarano and

Appeldoorn, 2013). These findings imply potential LBSP impacts across very large temporal and

spatial scales, with very wide and persistent implications on coral reef benthic communities and

on reef-associated fauna. LBSP impacts (i.e., sewage pollution from human and animal sources)

were documented across the entire southwestern shelf in Puerto Rico, even in waters complying

with existing microbiological quality standards (Bonkosky et al., 2009). This points out at the

increasing spatial scale of chronic LBSP impacts across multiple coral reef systems and at the

potentially increasing turnover rates of reef communities. The lack of adequate controls of LBSP

across the region constitutes one of the most significant concerns regarding the conservation and

recovery of declining coral reef ecosystems.


3

Efforts are being currently developed to implement erosion and sedimentation controls across

watershed scales in southwestern Puerto Rico. But so far, these efforts have completely missed a

long-term ecological monitoring component to determine if current land-based efforts have had

any meaningful impacts on improving adjacent coral reef ecosystems. To obtain a spatial

synoptic understanding of coral reef health and the land-based factors affecting it a broad-scale

assessment of reef condition is needed.

This project conducted a large-scale assessment across 11 coral reefs located along a LBSP

gradient across the southwestern Puerto Rico shelf. Standardized random photo-transect

techniques were used across one to four different depth zones when present (5 m, 5-10 m, 10-15

m, 15-20 m), depending on the site, to characterize benthic communities and assess coral recruit

density. Fish community structure was also assessed across similar depth zones. In addition, a

snapshot characterization of water quality parameters was determined during each visit to study

sites to establish a baseline data bank. This approach provided critical baseline information

regarding the current status of benthic and fish communities across a LBSP stress gradient. This

information will allow designing a long-term ecological monitoring program specifically

targeted at addressing impacts from LBSP on reef ecosystems, as well as the impacts of best

management practices (BMPs) implemented on selected watersheds.

The objectives of this project were to:

(1) Provide a large-scale spatial characterization over a variety of reef locations and conditions of

the health of coral reef benthic communities, coral recruit assemblages, and reef fish

communities.

(2) Obtain a snapshot characterization of water quality across each sampling site.

(3) Infer possible causative factors underlying observed spatial patterns in benthic and fish

community structure.


4

This design allowed us testing multiple null hypotheses regarding spatial patterns in benthic

community structure, coral recruit assemblages, and reef fish community structure, and in water

quality parameters among geographic locations (inshore, mid-shelf, outer shelf), reefs (within

each individual geographic region), and among depth zones. This information will be made

available to decision-makers at the PR Department of Natural and Environmental Resources

(PRDNER) and at the National Oceanic and Atmospheric Administration (NOAA) to help

implement future management actions.

II. METHODS

Study sites – A total of eleven coral reefs were studied, subdivided in three different geographic

locations (Figures 1-2): Inshore Reefs (<4 km) – Cayo Ratones (RAT), Punta Ostiones (OST),

Punta Lamela (LAM), Punta Guaniquilla (GUA), Arrecife Enmedio (EME); Mid-Shelf Reefs (4-

8 km) – Bajo Resuello (RES), Arrecife Corona del Norte (CON), Arrecife El Ron (RON); and

Outer Shelf Reefs (8-15 km) – Arrecife El Negro (NEG); Arrecife Papa San-Tourmaline (PPS),

Bajo Gallardo (GAL). Sampling was completed during July 13-30, 2014. Six of the above sites

(55%) were located within Natural Reserves managed by PRDNER, including the Marine

Extension of Finca Belvedere Natural Reserve, off Puerto Real (OST), the Marine Extension of

Punta Guaniquilla Natural Reserve, off Boquerón Bay (GUA), Cayo Ratones Natural Reserve

(RAT), and Tourmaline Natural Reserve (RON, CON, PPS). From these, there is a 6 month-

seasonal fishing closure at Tourmaline Natural Reserve. Fishing remains open across all other

reserves.

Benthic communities – Coral species richness, colony abundance, percentage cover of all major

benthic components (i.e., corals, algal functional groups, sponges, other macroinvertebrates), and

coral species diversity and evenness indices were documented. When possible, sources of recent

mortality in coral colonies were identified. Data was collected from randomly-placed 10 m-long

photo-transects from four different depth zones when present (<5 m, 5-10 m, 10-15 m, 15-20 m).

Briefly, a total of 5 to 15 replicate transects per depth zone were obtained using 5 non-

overlapping, high-resolution digital images from a roughly 1.0 x 0.7 m quadrat per transect at

fixed intervals (1, 3, 5, 7, and 9 m along each transect). This provided a total of 25-75 images per


5

depth zone/reef. A digital grid of 48 regularly-distributed dots was projected over each image for

analysis of percent coverage of benthic components. A 3-way permutational analysis of variance

(PERMANOVA) design (geographic location, reef site, depth) was used to test hypotheses

regarding spatial and variation patterns in benthic community parameters and benthic community

structure. Principal component ordination (PCO) was determined in PRIMER-e 6.1.16 +

PERMANOVA 1.06 (Plymouth Marine Laboratory, UK) to test for indicator benthic categories

of spatial clustering patterns of coral reef benthic communities (Clarke and Warwick, 2001;

Anderson et al., 2008).

Coral recruit communities – Coral recruit species richness and colony abundance were

documented following a similar sampling design. Data was collected from five replicate,

randomly-placed 10 m-long photo-transects from four different depth zones when present (<5 m,

5-10 m, 10-15 m, 15-20 m). Data was obtained using 5 non-overlapping, high-resolution digital

images from a roughly 0.40 x 0.27 m quadrat per transect at fixed intervals (1, 3, 5, 7, and 9 m

along each transect). This provided a total of 25 images per depth zone/reef. All coral recruits

were counted and identified to the lowest taxon possible. A 3-way permutational analysis of

variance (PERMANOVA) design (geographic location, reef site, depth) was used to test

hypotheses regarding spatial and variation patterns in coral recruit community structure. PCO

was determined to test for indicator coral recruit species of spatial clustering patterns of coral

recruit communities (Clarke and Warwick, 2001; Anderson et al., 2008).

Fish communities – Fish communities were sampled within 7.5 m-radius circular transects.

Sampling was conducted at 10-min time intervals and through similar depth zones, with a total of

4-6 replicate censuses/depth zone. A 3-way factorial design as above (geographic location, reef

site, depth) was used to test hypotheses regarding spatial and variation patterns in fish

community parameters and fish community structure. Fork lengths (FL) were converted to

weight (Bohnsack and Harper 1988, Bohnsack unpub. data). Abundance within large fish

schools were estimated by 10s, 20s, 50s or 100s. Fish species richness, species diversity and

evenness, abundance and biomass were calculated. Multivariate SIMPER analysis of indicator

species (Clarke and Warwick, 2001) and PCO was used to test for indicator species of spatial

clustering patterns.


6

FIGURE 1. Selected sampling sites across the Southwestern PR insular shelf. Sites were subdivided in three

different geographic locations: Inshore Reefs (<4 km) – Cayo Ratones (RAT), Punta Ostiones

(OST), Punta Lamela (LAM), Punta Guaniquilla (GUA), Arrecife Enmedio (EME); Mid-Shelf

Reefs (4-8 km) – Bajo Resuello (RES), Arrecife Corona del Norte (CON), Arrecife El Ron (RON);

and Outer Shelf Reefs (8-15 km) – Arrecife El Negro (NEG), Arrecife Papa San (PPS), Bajo

Gallardo (GAL). Eleven different protected areas surround the study sites: BEB= Boquerón State

Forest; CRNWR= Cabo Rojo National Wildlife Refuge (USFWS), EMRNFB= Finca Belvedere

Natural Reserve Marine Extension; EMRNPG= Punta Guaniquilla Natural Reserve Marine

Extension; EMBEB= Boquerón State Forest Marine Extension; RVSIAB= Boquerón’s Iris

Alameda Wildlife Refuge; RNAT= Arrecife Tourmaline Natural Reserve; RNCR= Cayo Ratones

Natural Reserve; RNFB= Finca Belvedere Natural Reserve; RNLJ= Laguna Joyuda Natural

Reserve; RNPG= Punta Guaniquilla Natural Reserve.


7

FIGURE 2. Selected sampling sites across the Southwestern PR insular shelf. Sites were

subdivided in three different geographic locations: Inshore Reefs (<4 km) – Cayo

Ratones (RAT), Punta Ostiones (OST), Punta Lamela (LAM), Punta Guaniquilla

(GUA), Arrecife Enmedio (EME); Mid-Shelf Reefs (4-8 km) – Bajo Resuello

(RES), Arrecife Corona del Norte (CON), Arrecife El Ron (RON); and Outer

Shelf Reefs (8-15 km) – Arrecife El Negro (NEG), Arrecife Papa San (PPS), Bajo

Gallardo (GAL).

Water quality parameters – Water quality parameters were documented at each site, including

temperature, salinity, conductivity and dissolved oxygen concentration using a YSI85 data

logger. Turbidity was measured using a LaMotte turbidimeter. Chlorophyll-a concentration and

optical brighteners concentration was documented using a Turner fluorometer. Phosphate (PO4),

ammonium (NH3), and ionized ammonium (NH4+) concentrations were determined using a

Smart V2 spectrophotometer (LaMotte). Triplicate samples were obtained per site at 30 cm

below surface. Water quality data was analyzed following a two-way PERMANOVA with

geographic location and reef site as main factors to test null hypotheses regarding spatial


8

variation patterns in water quality. PCO was determined used to test for indicator water quality

parameters of spatial LBSP patterns (Clarke and Warwick, 2001; Anderson et al., 2008).

Multivariate RELATE routine from PRIMER-e 6.1.16 was used to test for correlations of water

quality parameters with benthic community structure, coral recruit community structure, and fish

community structure. Multivariate LINKTREE test was conducted following BEST routine and

independent BIOENV and BVSTEP stepwise regression procedures to determine which

individual or combined water quality parameters better explained observed spatial patterns in

biological community structure for benthic, coral recruit, and fish communities.

Cross-shelf spatial patterns of LBSP

Arc-View 10.2 (ESRI) was used to address cross-shelf spatial patterns of water quality patterns

based on the preliminary sampling of multiple water quality parameters.

Site

JOY OST LAM GUA RAT EME RES CON RON NEG PPS GAL

SS

T

28.6

28.8

29.0

29.2

29.4

29.6

29.8

30.0

30.2

30.4

FIGURE 3. Sea surface temperature across sites. Red circles= Inshore sites; Green circles=

Mid-shelf sites; Black circles= Outer shelf sites. Mean±95% confidence interval.


9

III. RESULTS

a. Water quality

Sea surface temperature (SST) was significantly higher and had a higher fluctuation at shallower

inshore sites (29.6°C to 30.2°C), than at mid-shelf sites (29.6°C to 29.7°C), and outer shelf sites

(28.7°C to 29.0°C) (Figure 3). Salinity ranged from 35.4 to 35.7 ppt across inshore sites, from

35.2 to 35.7 ppt across mid-shelf sites, and from 35.1 to 35.7 ppt across outer shelf sites (Figure

4). Conductivity showed a spatial gradient with higher mean values across inshore sites (58.0 to

59.4 mS), followed by mid-shelf sites (58.3 to 59.0 mS), and outer shelf sites (57.2 to 58.0 mS)

(Figure 5). The observed pattern suggest a relationship between warmer SST and higher salinity

across inshore sites, in comparison to outer shelf localities. This pattern could be the result of

very complex cross-shelf tidal circulation as well as by regional circulation factors.

Dissolved oxygen concentration also showed a significant increase with increased distance from

known polluted coastal reef communities (Figure 6). Inshore reef sites showed a range from 3.8

to 4.3 mg/L. These values became particularly low under the influence of turbid ebbing tides as

previously documented by Bonkosky et al. (2009). Dissolved oxygen concentration fluctuated

from 3.8 to 4.5 mg/L across mid-shelf sites, and from 5.5 to 5.8 mg/L across outer shelf sites.

Mean turbidity showed an opposite trend, with higher mean values across inshore sites, which

showed a range from 1.0 to 3.8 NTU (Figure 7). Mid-shelf sites averaged 0.9 to 1.0 NTU, and

outer shelf sites 0.4 to 0.9 NTU. Dissolved oxygen and turbidity patterns show often complex

spatial and temporal variability across the western shelf due to complex circulation patterns.

Chlorophyll-a concentration ranged from 1.50 to 2.69 μg/L across inshore sites, from 1.83 to

2.31 μg/L across mid-shelf sites, and from 1.89 to 2.29 μg/L across outer shelf sites (Figure 8).

Upper mean values were documented across Punta Ostiones, Cayo Ratones and Bajo Enmedio,

all across the inshore sites. Optical brighteners concentration (OABs) ranged from 17 to 32 ppm

across inshore sites, from 22 to 28 ppm across mid-shelf sites, and from 17 to 25 ppm across

outer shelf sites. OABs concentration showed large fluctuations as a result of complex ocean

circulation patterns across the shelf. Nonetheless, higher OABs concentrations were observed at

Joyuda and at Bajo Enmedio, off Boquerón Bay mouth. Both sites are known to receive raw

sewage and gray water discharges directly to coastal waters. OABs concentrations were also

fairly high at Arrecife El Ron and at Corona del Norte mid-shelf sites. These also receive Río


10

Site


Sa

linity (

pp

t)

35.0

35.1

35.2

35.3

35.4

35.5

35.6

35.7

35.8

Inshore Mid Offshore

FIGURE 4. Salinity across sites. Red circles= Inshore sites; Green circles= Mid-shelf sites;

Black circles= Outer shelf sites. Mean±95% confidence interval.

Site


Cond

uctivity (

mS

)

56.5

57.0

57.5

58.0

58.5

59.0

59.5

60.0


FIGURE 5. Salinity across sites. Red circles= Inshore sites; Green circles= Mid-shelf sites;



11

Site


Dis

solv

ed

oxygen

(m

g/L

)

3.5

4.0

4.5

5.0

5.5

6.0


FIGURE 6. Dissolved oxygen across sites. Red circles= Inshore sites; Green circles= Mid-

shelf sites; Black circles= Outer shelf sites. Mean±95% confidence interval.

Site


Turb

idity (

NT

U)

0

1

2

3

4

5


FIGURE 7. Turbidity across sites. Red circles= Inshore sites; Green circles= Mid-shelf sites;



12

Site


Chlo

rophyll-

a (

ug/L

)

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0


FIGURE 8. Chlorophyll-a concentration across sites. Red circles= Inshore sites; Green

circles= Mid-shelf sites; Black circles= Outer shelf sites. Mean±95% confidence

interval.

Guanajibo effluents during ebbing tides through the Guanajibo Channel.

Phosphate (PO4) concentrations resulted higher within inshore sites, ranging from 2.2 to 4.4 μM

(Figure 10). PO4 concentrations ranged from 2.5 to 3.0 μM across inshore sites, and from 0.8 to

1.9 μM across outer shelf sites. This reflects a concentration gradient with increasing distance.

Ammonium concentration (NH3) showed large spatial variability, with inshore sites ranging from

25 to 264 μM (Figure 11). Mid-shelf sites ranged from 22 to 133 μM, and outer shelf sites

ranged from 15 to 16 μM. Bajo Enmedio (264 μM), Punta Guaniquilla (136 μM), and Bajo

Resuello (133 μM), which are located just off Boquerón Bay and are known to receive recurrent

raw sewage illegal discharges and poorly-treated sewage effluents from a malfunctioning

treatment facility from Boquerón Bay, showed the highest NH3 concentrations. NH3

concentration at Punta Lamela, located just off Puerto Real, showed also a concentration of 94

μM, which is also considered very high.


13

Site


OA

Bs (

pp

m)

10

15

20

25

30

35

40


FIGURE 9. Optical brighteners (OABs) concentration across sites. Red circles= Inshore sites;

Green circles= Mid-shelf sites; Black circles= Outer shelf sites. Mean±95%

confidence interval.

Site


PO

4 (

uM

)

0

1

2

3

4

5

6

7


FIGURE 10. Phosphate (PO4) concentration across sites. Red circles= Inshore sites; Green


interval.


14

Site


NH

3 (

uM

)

0

50

100

150

200

250

300

350


FIGURE 11. Ammonium (NH3) concentration across sites. Red circles= Inshore sites; Green


interval.

Site


NH

4

+ (

uM

)

0

50

100

150

200

250

300

350


FIGURE 12. Ionized ammonium (NH4+) concentration across sites. Red circles= Inshore sites;




15

r2=0.3869, p=0.0308f = 3.90+0.09*x

Distance from LBSP (km)

0 2 4 6 8 10 12 14 16 18

Dis

solv

ed

oxygen

(m

g/L

)

2

3

4

5

6

7

Dissolved oxygen

Regression line

95% Confidence Band

FIGURE 13. Linear regression analysis between dissolved oxygen concentration and distance

from the known pollution centers along the coast.

Ionized ammonium concentration (NH4+) followed similar spatial patterns, with inshore sites

ranging from 24 to 260 μM (Figure 12). Mid-shelf sites ranged from 21 to 131 μM, and outer

shelf sites ranged from 15 to 16 μM. Bajo Enmedio (260 μM), Punta Guaniquilla (134 μM), and

Bajo Resuello (131 μM) also showed the highest NH4+ concentrations. NH4

+ concentration at

Punta Lamela, located just off Puerto Real, showed also a concentration of 93 μM, which is also

considered very high. These observations are highly consistent with previous observations of

significant sewage pollution impacts across the western Puerto Rican shelf (Bonkosky et al.,

2009; Hernández-Delgado et al., 2010).

A linear regression analysis showed a significant (r2=0.3869; p=0.0308) increase in dissolved

oxygen concentration with increasing distance from known pollution centers along the coast

(Figure 13). Also, water turbidity showed a highly significant decline (r2=0.7119; p=0.0006)

with increasing distance from the coast (Figure 14). No significant spatial gradient was found

with the chlorophyll-a concentration (Figure 15) and OABs concentration (Figure 16) and

increasing distance from LBSP. Phosphate (PO4) showed a significant decline (r2=0.3952;

p=0.0286) with increasing distance from LBSP (Figure 17). There was a significant (r2=0.4961;


16

r2=0.7119, p=0.0006

f = 3.13-0.21*x


0 2 4 6 8 10 12 14 16 18

Turb

idity (

NT

U)

0

1

2

3

4

5

Turbidity

Regression line

95% Confidence Band

FIGURE 14. Linear regression analysis between turbidity and distance from the known

pollution centers along the coast.

r2=0.2742, p=0.0806

f = 2.50-0.04*x


0 2 4 6 8 10 12 14 16 18

Ch

loro

ph

yll-

a (

ug

/L)

1.0

1.5

2.0

2.5

3.0

3.5

Chlorophyll-a

Regression line

95% Confidence Band

FIGURE 15. Linear regression analysis between chlorophyll-a concentration and distance from

the known pollution centers along the coast.


17

r2=0.0199, p=0.6617f = 24.74-0.15*x


0 2 4 6 8 10 12 14 16 18

OA

Bs (

pp

m)

10

15

20

25

30

35

OABs

Regression line

95% Confidence Band

FIGURE 16. Linear regression analysis between optical brighteners (OABs) concentration and

distance from the known pollution centers along the coast.

r2=0.3952, p=0.0286

f = 3.64-0.15*x


0 2 4 6 8 10 12 14 16 18

PO

4 (

uM

)

0

1

2

3

4

5

6

PO4

Regression line

95% Confidence Band

FIGURE 17. Linear regression analysis between phosphate (PO4) concentration and distance

from the known pollution centers along the coast.


18

p=0.0458) non-linear relationship between NH3 (Figure 18) and NH4+ (Figure 19) with

increasing distance from LBSP. There was a significant non-linear decline (r2=0.6174;

p=0.0133) in dissolved oxygen concentration with increasing turbidity (Figure 20). Chlorophyll-

a concentration showed a significant linear increase (r2=0.5440; p=0.0062) with increasing

turbidity (Figure 21).

There was also a significant linear increase (r2=0.6579; p=0.0014) in PO4 concentration with

increasing turbidity (Figure 22), and a significant non-linear increase (r2=0.5121; p=0.0396) in

chlorophyll-a concentration with increasing PO4 concentration (Figure 23). In addition, there

was a significant linear decline (r2=0.5364; p=0.0068) in dissolved oxygen concentration with

increasing PO4 concentration (Figure 24).

There was a highly significant (p<0.0001) spatial gradient of LBSP (Table 1) which produced

four different clusters. A first large cluster was composed of inshore sites JOY, OST, RAT,

LAM, and mid-shelf sites CON, RON (Figure 25). OST and RAT were largely explained by

high chlorophyll-a concentration. JOY was largely explained by high NTU and OABs. LAM and

EME were explained by high PO4, sea surface temperature (SST) and conductivity. CON and

RON were explained by high salinity. EME constituted a second individual cluster. A third

cluster was composed of GUA and RES and was explained by high NH4+. A final cluster was

composed by NEG, PPS, GAL and was explained by higher dissolved oxygen concentration.

This model explained 74.3% of the spatial variation observed in water quality parameters across

the entire shelf.

Cross-shelf spatial patterns of LBSP

GIS-based analysis was used to determine cross-shelf spatial patterns in selected water quality

parameters. Salinity showed slightly higher values across northern inshore and mid-shelf sites

(Figure 26). Dissolved oxygen concentration showed lower values across inshore RAT site and

mid-shelf sites CON, RON (Figure 27). All other inshore sites showed moderately low values.

Higher values were observed at outer shelf sites. With exception of inshore site GUA, turbidity

was higher across all inshore sites and moderate across mid-shelf sites (Figure 28). Turbidity

was lower across outer shelf sites. Chlorophyll-a was higher across inshore sites, and moderate


19

r2=0.4961, p=0.0458

f=162.64*exp(-.5*((x-4.6)/2.39)^2)


0 2 4 6 8 10 12 14 16 18

NH

3 (

uM

)

0

50

100

150

200

250

300

NH3

Regression line

95% Confidence Band

FIGURE 18. Non-linear regression analysis between ammonium (NH3) concentration and

distance from the known pollution centers along the coast.

r2=0.4961, p=0.0458

f=160.17*exp(-.5*((x-4.6)/2.39)^2)


0 2 4 6 8 10 12 14 16 18

NH

4

+ (

uM

)

0

50

100

150

200

250

300

NH4

+

Regression line

95% Confidence Band

FIGURE 19. Non-linear regression analysis between ionized ammonium (NH4+) concentration

and distance from the known pollution centers along the coast.


20

r2=0.6174, p=0.0133

f=3.88*exp(0.13/(x-0.06))

Turbidity (NTU)

1 2 3 4

Dis

solv

ed

oxygen

(m

g/L

)

1

2

3

4

5

6

Dissolved oxygen

Regression line

95% Confidence Band

FIGURE 20. Non-linear regression analysis between dissolved oxygen concentration and

turbidity.

r2=0.5440, p=0.0062f = 1.83+0.23*x

Turbidity (NTU)

0 1 2 3 4

Ch

loro

ph

yll-

a (

ug

/L)

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

3.2

3.4

Chlorophyll-a

Regression line

95% Confidence Band

FIGURE 21. Linear regression analysis between chlorophyll-a concentration and turbidity.


21

r2=0.6579, p=0.0014f = 1.37+0.75*x

Turbidity (NTU)

0 1 2 3 4

PO

4 (

uM

)

0

1

2

3

4

5

6

PO4

Regression line

95% Confidence Band

FIGURE 22. Linear regression analysis between phosphate (PO4) concentration and turbidity.

r2=0.5121, p=0.0396

f=2.48-0.52*x+0.14*x^2

PO4 (uM)

0 1 2 3 4 5

Ch

loro

ph

yll-

a (

ug

/L)

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

Chlorophyll-a

Regression line

95% Confidence Band

FIGURE 23. Non-linear regression analysis between cholorphyll-a concentration and

phosphate (PO4) concentration.


22

r2=0.5364, p=0.0068

f = 5.75-0.47*x

PO4 (uM)

0 1 2 3 4 5

Dis

so

lve

d o

xyge

n (

mg/L

)

2

3

4

5

6

7

Dissolved oxygen

Regression line

95% Confidence Band

FIGURE 24. Linear regression analysis between dissolved oxygen concentration and phosphate

(PO4) concentration.

TABLE 1. Summary of a two-way PERMANOVA test of sampling sites clustering patterns

based on water quality parameters.

Variable d.f. Pseudo-

F p

Geographic location 2,33 13.67 <0.0001 Site 11,24 13.66 <0.0001 Location x Site 11,24 13.66 <0.0001


23

FIGURE 25. Principal component ordination (PCO) plot of the LBSP stress gradient across

study sites. Based on Euclidean distance of Log10-transformed water quality

parameters. NH3 data was removed from the matrix due to co-linearity with NH4+

data. Total variation= 74.3%.

-4 -2 0 2 4

PCO1 (56.1% of total variation)

-4

-2

0

2P

CO

2 (

18

.2%

of

tota

l va

ria

tio

n)

Distance4

JOY

RAT

OST

LAM

GUA

EME

RES

CONRON

NEG

PPS

GAL

TC

Sal

Cond

D.O

NTU

Chl-a

OABs

PO4

NH4+


24

FIGURE 26. Cross-shelf spatial patterns of salinity.


25

FIGURE x. Cross-shelf spatial patterns of dissolved oxygen.


26

FIGURE 28. Cross-shelf spatial patterns of turbidity.


27

across northern inshore, mid-shelf and offshore sites, suggesting potential influences from north

south surface circulation from Mayagüez Bay (Figure 29). OABs concentrations showed higher

and moderate concentrations across inshore reef sites, particularly adjacent to Boquerón Bay and

Playa Joyuda (Figure 30).

PO4 concentrations were significantly higher across all inshore sites in comparison to mid-shelf

and outer sites, suggesting a direct linear relationship with proximity to LBSP (Figure 31). NH4+

concentrations were also concerning and higher across inshore sites, particularly in those

adjacent to Boquerón Bay and Puerto Real Bay (Figure 32). Observed spatial trends suggest five

possible terrestrial influences of pollution. The most critical one appears to be raw sewage

effluents and septic tank infiltration from multiple non-point sources across Boquerón Bay, plus

poorly-treated discharges from four different package sewage treatment plants. The second

source of LBSP is raw sewage pollution from Puerto Real Bay and septic tank infiltration from

private houses from Puerto Real and Punta Lamela. A third potential source of human-borne

pollution is the multiple non-point sources of raw sewage, gray waters, and septic tank

infiltration from Playa Joyuda. A fourth significant source of pollution can come from the Río

Guanajibo outlet. Its effluents move to the west, south-west, across Las Coronas coral reef

system across the Guanajibo Channel. A final potential source of LBSP is Mayagüez Bay. These

findings are very consistent with Bonkosky et al. (2009) and suggest the complex nature of water

circulation pattern and the network of terrestrial-marine connectivity across the southwestern

region.

Discussion of water quality observations

This study had the limitation that sampling was conducted only once, so it must be interpreted

with caution. However, a key concern associated to sewage and eutrophication impacts from

LBSP across the western Puerto Rican shelf in this study was the elevated mean chlorophyll-a

concentration range found in this study across the shelf (1.50-2.69 μg/L). These values were up

to 4-8 times higher than the recommended concentration for coral reef waters (0.3-0.5 μg/L)

(Lapointe and Clark, 1992; Otero, 2009). Recent studies conducted at severely eutrophied Vega

Baja beach, in northern Puerto Rico, showed chlorophyll-a concentrations of up to 17-83 times

higher than recommended limits (Díaz-Ortega and Hernández-Delgado, 2014).


28

FIGURE 29. Cross-shelf spatial patterns of chlorophyll-a concentration.


29

FIGURE 30. Cross-shelf spatial patterns of optical brighteners (OABs) concentration.


30

FIGURE 31. Cross-shelf spatial patterns of phosphate (PO4) concentration.


31

FIGURE 32. Cross-shelf spatial patterns of ionized ammonium (NH4+) concentration.


32

OABs ranged in this study from 16 to 32 ppm across the shelf. These values were similar to

those obtained by Díaz-Ortega and Hernández-Delgado (2014) under moderately polluted reef

zones, where highly polluted areas showed OAB concentrations ranging from 50 to 200 ppm.

Our findings indicated potentially moderate human pollution across shelf-wide scales in western

Puerto Rico. PO4 concentrations ranged from 0.8 to 4.4 μM across the shelf in this study.

Lapointe and Clark (1992) suggested that PO4 concentration on coral reef habitats should not

exceed 0.3 μM. Lapointe and Matzie (1996) suggested that any concentration above 1.5 μM

would be too high for coral reefs. Our findings suggest that concentrations observed were from

1.7 to 13.7 times higher than the recommended limits for coral reefs. Further, 10 out of the 12

sampled sites (83%) showed PO4 concentrations that were from 0.3 to 1.9 times higher than the

concentration considered unsustainable for coral reefs.

NH4+ concentrations in this study ranged from 15 to 260 μM. Lapointe and Clark (1992)

suggested that NH4+ concentrations for coral reefs should not exceed 0.1 μM, and that any

concentration above 24 μM were deemed as too high. Our findings are highly concerning as

observed NH4+ concentrations were from 149 to 2,590 times higher than recommended limits for

coral reefs. Eight out the twelve sampled sites (75%) showed NH4+ concentrations exceeding

dangerous NH4+ concentrations for coral reefs from 0.02 to 9.8 times.

Regression analyses showed that several water quality indicator parameters showed significant

gradients with increasing distance from LBSP. Also, there was a significant relationship among

turbidity, PO4, chlorophyll-a, and dissolved oxygen concentration, implying that increasing water

quality degradation significantly affects multiple water quality parameters, therefore adversely

impacting coral reefs. In addition, PCO analysis showed a definite shelf-wide LBSP spatial

gradient. Although this study just provided a snapshot view of water quality across the western

Puerto Rico shelf, results were concerning as critical water quality parameters resulted

significantly higher than recommended limits for sustaining coral reef health. These results

suggest that human-driven LBSP across the western Puerto Rico shelf is highly significant, it is a

large-scale, chronic phenomenon, and deserves full long-term monitoring across large spatial and

temporal scales.


33

B. Coral reef benthic communities

Coral reef benthic community structure showed a highly significant variation among all

geographic zones, sites and depth zones, suggesting a strong effect associated to the LBSP

gradient and to depth (Table 2). A PCO analysis showed that coral reef benthic communities

along the western Puerto Rican shelf were clustered in three different groups characterized by

different dominant coral taxa (Figure 33). A first group was entirely composed of inshore reef

sites RAT, OST, LAM and GUA, and was explained by a combination of green and red

macroalgae. The second cluster was composed by EME and RES. The former was the farthest

offshore coral reef among all inshore sites and the latter was the closest mid-shelf site. Both are

located offshore Boquerón Bay (Figures 1-2). EME was dominated by octocorals

Pseudoplexaura spp. and by Eunicea laxispica. The third cluster was composed by mid-shelf

CON, RON sites and by outer shelf sites NEG, PPS, GAL. PPS was explained by scleractinian

coral Montastraea cavernosa and octocoral Gorgonia ventalina. RON was explained by

Orbicella franksi and by dominant cyanobacteria Lyngbya spp. The remaining sites were

explained by scleractinian Undaria agaricites. Percent live coral cover showed a significant

increase with increasing distance from LBSP (Figures 34-35).

TABLE 2. Summary of a three-way PERMANOVA test of benthic community structure


F p

Geographic location 2,254 22.15 <0.0001 Site 10,246 14.04 <0.0001 Depth 3,253 8.02 <0.0001 Location x Site 10,246 14.04 <0.0001 Location x Depth 8,248 9.97 <0.0001 Site x Depth 22,234 8.78 <0.0001 Location x Site x Depth 22,234 8.78 <0.0001


34

FIGURE 33. Principal component ordination (PCO) of benthic community structure among

study sites. Clusters based on a 60% similarity cutoff level. Vectors based on a

0.60 correlation. This model explained 56.4% of the observed variation.

Mean Scleractinian + Hydrocoral species richness increased with increasing distance from the

shore, and averaged 2.96/transect across inshore sites, with the lowest value across the 5-10 m

depth zone of EME (0.9/transect) and the highest at the <5 m depth zone of GUA (4.0/transect)

(Figure 34). Mean species richness across mid-shelf sites averaged 6.26/transect, with the lowest

value across the <5 m depth zone of CON (3.6/transect) and the highest at the 10-15 m depth

zone of RON (9.5/transect). Mean species richness across outer shelf sites averaged

7.13/transect, with the lowest value across the <5 m depth zone of GAL (4.3/transect) and the

highest at the 10-15 m depth zone of PPS (8.73/transect).

-30 -20 -10 0 10 20 30


-30

-20

-10

0

10

20P

CO

2 (

21.6

% o

f to

tal va

riation)

ZoneIn

Mid

Out

Similarity60

OST

LAM

GUA

RAT

EME

RES

CON

RON

NEG

PPS

GAL

Mcav

Ofra

Uaga

Elax

Gven

Psdp

Pspt

Gmac

RedMac

Lyng


35

Location x Site x Depth

OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

Sp

ecie

s r

ich

ne

ss

0

2

4

6

8

10

12

14

FIGURE 34. Sleractinian + Hydrocoral species richness across sites. Red circles= Inshore sites;



Mean coral species diversity (H’n) also increased with increasing distance from the shore, and

averaged 0.9156/transect across inshore sites, with the lowest value across the 5-10 m depth zone

of EME (0.2367) and the highest at the <5 m depth zone of GUA (1.2648) (Figure 35). H’n

across mid-shelf sites averaged 1.6626/transect, with the lowest value across the <5 m depth

zone of CON (1.1087) and the highest at the 10-15 m depth zone of RON (2.0741). H’n across

outer shelf sites averaged 1.7829/transect, with the lowest value across the <5 m depth zone of

GAL (1.2519) and the highest at the 10-15 m depth zone of PPS (2.0595).

Mean coral species evenness (J’n) also increased with increasing distance from the shore, and

averaged 0.7543/transect across inshore sites, with the lowest value across the 5-10 m depth zone

of EME (0.2843) and the highest at the <5 m depth zone of LAM (0.9531) (Figure 36). J’n

across mid-shelf sites averaged 0.9302/transect, with the lowest value across the <5 m depth

zone of CON (0.8302) and the highest at the 10-15 m depth zone of CON (0.9641). J’n across

outer shelf sites averaged 0.9345/transect, with the lowest value across the <5 m depth zone of

GAL (0.8686) and the highest at the 10-15 m depth zone of PPS (0.9707).


36


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

H'n

0.0

0.5

1.0

1.5

2.0

2.5

FIGURE 35. Sleractinian + Hydrocoral species diversity index (H’n) across sites. Red circles=

Inshore sites; Green circles= Mid-shelf sites; Black circles= Outer shelf sites.

Mean±95% confidence interval.


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

J'n

0.0

0.2

0.4

0.6

0.8

1.0

FIGURE 36. Sleractinian + Hydrocoral species diversity index (H’n) across sites. Red circles=




37

FIGURE 37. Cross-shelf spatial patterns of percent living coral cover.


38


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% C

ora

l

0

5

10

15

20

25

30

FIGURE 38. Percent living coral cover across sites. Red circles= Inshore sites; Green circles=


Mean percent coral cover increased with increasing distance from the shore (Figure 37), and

averaged 7.8% across inshore sites, with the lowest value across the 5-10 m depth zone of EME

(0.8%) and the highest at GUA (10.8%) (Figure 38). Mean percent cover across mid-shelf sites

averaged 15.1%, with the lowest value across the 10-15 m depth zone of CON (11.2%) and the

highest at the 5-10 m depth zone of RON (22.5%). Mean percent cover across outer shelf sites

averaged 15.8%, with the lowest value across the 15-20 m depth zone of PPS (11.2%) and the

highest at the 5-10 m depth zone of NEG (23.9%).

The dominant scleractinian coral at OST was Siderastrea sidera, while Porites porites was the

dominant coral at LAM, P. astreoides at GUA, and Orbicella faveolata at RAT and the shallow

zone of EME (Figure 39). The dominant coral at EME deeper zone was also O. faveolata, while

P. astreoides was dominant at RES shallow and middle depth zones. Orbicella faveolata was the

most common coral across the deeper zone at RES. Porites astreoides was the most common

coral across all depth zones at CON, while O. faveolata was the most abundant at the shallow

and middle depth zones a RON.


39


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% C

ora

ls

0

2

4

6

8

10

12

14

P. astreoides

P. porites

P. divaricata

P. furcata

O. faveolata

O. annularis

O. franksi

M. cavernosa

S. siderea

S. radians

A. palmata

U. agaricites

U. tenuifolia

C. natans

D. cylindrus

P. strigosa

D. labyrinthiformis

M. alcicornis

Others (22)

FIGURE 39. Percent relative cover of the most abundant Scleractinian corals + Hydrocorals.

Montastraea cavernosa was dominant across deeper zones at RON. Porites astreoides was the

most abundant species across the shallower and deeper zones at NEG, while P. porites was the

dominant scleractinian at the middle depth zones at NEG. Porites astreoides was the most

common species at the 10-15 m zone and M. cavernosa at the 15-20 m depth zone at PPS.

Acropora palmata was the most common species at the shallower and middle depth zone at

GAL, while O. annularis was the dominant scleractinian species across the deeper zone at GAL.


40


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% S

. sid

ere

a

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

FIGURE 40. Percent cover of Siderastrea siderea across sites. Red circles= Inshore sites;



Several scleractian coral species show particular spatial patterns that may evidence chronic

impacts by LBSP, including high turbidity, high sedimentation rates and eutrophication. For

instance, Siderastrea siderea, which is a coral commonly found across turbid conditions, was

more frequently documented across some of the inshore sites and across RES, which was the

most adjacent to shore of the mid-shelf sites (Figure 40). These sites resulted with higher

turbidity and dissolved nutrient levels. Montastraea cavernosa was more abundant at the deeper

zone of RON and at the middle depth and deeper zone of RES (Figure 41). It was also abundant

at PPS. Undaria agaricites, in contrast, was absent from inshore sites and was more abundant at

NEG, particularly on deeper zones (Figure 42). Threatened Elkhorn coral, Acropora palmata,

was nearly absent across all sites, with the exception of the shallower and mid-shelf zones of

Arrecife Gallardo (Figure 43). Threatened Orbicella annularis was overall more abundant

across outer shelf reefs, particularly at NEG and at the deeper zone of GAL (Figure 44).

Threatened O. faveolata was more abundant across mid-shelf reefs, particularly at the deeper

zone of RES and at the shallow and middle zones of RON (Figure 45). Threatened Dendrogyra

cylindrus was overall rare across range, but more common across outer shelf reefs (Figure 46).


41


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% M

. ca

ve

rno

sa

0

1

2

3

4

5

FIGURE 41. Percent cover of Montastraea cavernosa across sites. Red circles= Inshore sites;



Several scleractian coral species show particular spatial patterns that may evidence chronic

impacts by LBSP, including high turbidity, high sedimentation rates and eutrophication. For

instance, Siderastrea siderea, which is a coral commonly found across turbid conditions, was

more frequently documented across some of the inshore sites and across RES, which was the

most adjacent to shore of the mid-shelf sites (Figure 37). These sites resulted with higher

turbidity and dissolved nutrient levels. Montastraea cavernosa was more abundant at the deeper

zone of RON and at the middle depth and deeper zone of RES (Figure 38). It was also abundant

at PPS. Undaria agaricites, in contrast, was absent from inshore sites and was more abundant at

NEG, particularly on deeper zones (Figure 39). Threatened Elkhorn coral, Acropora palmata,

was nearly absent across all sites, with the exception of the shallower and mid-shelf zones of

Arrecife Gallardo (Figure 40). Threatened Orbicella annularis was overall more abundant

across outer shelf reefs, particularly at NEG and at the deeper zone of GAL (Figure 41).

Threatened O. faveolata was more abundant across mid-shelf reefs, particularly at the deeper

zone of RES and at the shallow and middle zones of RON (Figure 42). Threatened Dendrogyra

cylindrus was overall rare across range, but more common across outer shelf reefs (Figure 43).


42


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% U

. a

ga

ricite

s

0.0

0.5

1.0

1.5

2.0

2.5

3.0

FIGURE 42. Percent cover of Undaria agaricites across sites. Red circles= Inshore sites; Green


interval.


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% A

. p

alm

ata

0

2

4

6

8

FIGURE 43. Percent cover of E.S.A.-listed Acropora palmata across sites. Red circles=




43


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% O

. a

nn

ula

ris

0

1

2

3

4

FIGURE 44. Percent cover of E.S.A.-listed Orbicella annularis across sites. Red circles=




OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% O

. fa

ve

ola

ta

0

1

2

3

4

5

6

FIGURE 45. Percent cover of E.S.A.-listed Orbicella faveolata across sites. Red circles=




44


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% D

. cylin

dru

s

0.0

0.2

0.4

0.6

0.8

1.0

FIGURE 46. Percent cover of E.S.A.-listed Dendrogyra cylindrus across sites. Red circles=




OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% O

cto

co

ral

0

20

40

60

80

100

FIGURE 47. Percent octocoral cover across sites. Red circles= Inshore sites; Green circles=



45

Percent octocoral cover was significantly higher within shallower inshore sites subjected to

stronger influences by LBSP (Figure 47). Mean percent cover across inshore sites averaged

37.5%, with the lowest value across the 5-10 m depth zone of GUA (13.0%) and the two highest

values at the EME <5 m zone (57.1%) and its 5-10 m zone (71.5%). Highly degraded coral reefs

show a massive loss of living Scleractinian corals. Substrates were then largely occupied by

opportunist octocoral taxa. Mean percent octocoral cover across mid-shelf sites averaged 24.0%,

with the lowest value across the <5 m depth zone of CON (7.4%), the farthest mid-shelf site, and

the highest two values at the 5-10 m depth zone of RES (40.0%) and its 5-10 m depth zone

(47.8%). RES was the closest mid-shelf site, adjacent to polluted Boquerón Bay. Mean percent

cover across outer shelf sites averaged 15.1%, with the lowest value across the 5-10 m depth

zone of GAL (10.3%) and the highest at the 10-15 m depth zone of PPS (30.1%), followed by its

15-20 m depth zone across the upper vertical wall (25.6%). These results point out at the natural

dominance of octocoral communities at shallower inshore sites, as well as across deeper shelf

edge PPS site subjected to strong circulation. But also highlight the natural dominance of

octocorals on largely degraded habitats.

The most dominant octocoral species across the inshore sites was Pseudopterogorgia spp.,

followed by P. americana (Figure 48). These sites were highly influenced by LBSP.

Pseudopterogorgia spp. was also common across the mid-shelf RES mid-shelf site. They were

also common across other mid-shelf sites, but other species were also common, including

Gorgonia ventalina, Pseudoplexaura spp., Eunicea spp. Briareum asbestinum, and the

opportunist species Erythropodium caribbaeorum. These species were also common at the outer

shelf sites, but deeper habitats at mid-shelf NEG and wall habitats at PPS also showed abundant

red octocoral Icilligorgia shrammi. Black corals were also abundant at the PPS deeper wall (20-

30 m), but colonies were located just below surveyed transects. But Pseudopterogorgia spp.

(Figure 49), P. americana (Figure 50), and E. caribbaeorum (Figure 51) showed a significant

increase in abundance across highly disturbed sites, most of them inshore and adjacent to LBSP.

Percent sponge cover averaged 1.4% across inshore sites, with the lowest value at OST (0.06%)

and the highest at the EME 5-10 m zone (3.3%) (Figure 52). Mean percent sponge cover across

mid-shelf sites was 2.9%, with the lowest value at the <5 m depth zone of RON (0.6%), and the

highest at the 5-10 m depth zone of RES (6.2%). Mean percent sponge cover across the outer

shelf site was 2.2%, with the lowest value at the <5 m depth zone of GAL (0.5%), and the


46


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% O

cto

co

ral

0

10

20

30

40

Pseudopterogorgia spp.

Pspt. americana

Pspt. acerosa

Pspt. bipinnata

E. caribbaeorum

G. ventalina

Plexaura spp.

Pseudoplexaura spp.

Pspl. porosa

Pspl. flagellosa

Pspl. wagenaari

Plexaurella spp.

Pxla. nutans

Plx. homomalla

Plx. flexuosa

B. asbestinum

Eunicea spp.

E. tourneforti

I. shramii

Others

FIGURE 48. Percent relative cover of the most abundant octocorals.


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% P

se

ud

op

tero

go

rgia

sp

p.

0

5

10

15

20

25

FIGURE 49. Percent cover of Pseudopterogorgia spp. across sites. Red circles= Inshore sites;




47


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% P

. a

me

rica

na

0

2

4

6

8

10

12

14

16

FIGURE 50. Percent cover of Pseudopterogorgia americana across sites. Red circles= Inshore

sites; Green circles= Mid-shelf sites; Black circles= Outer shelf sites. Mean±95%

confidence interval


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% E

. caribbae

oru

m

0

2

4

6

8

10

FIGURE 51. Percent cover of Erythropodium caribbaeorum across sites. Red circles= Inshore




48


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% S

po

ng

e

0

2

4

6

8

10

FIGURE 52. Percent sponge cover across sites. Red circles= Inshore sites; Green circles= Mid-

shelf sites; Black circles= Outer shelf sites. Mean±95% confidence interval.

highest value at the 15-20 m deep upper wall segment of PPS (5.6%). This site in particular

showed also a very high abundance of the giant Barrel sponge, Xestospongia muta.

Percent macroalgal cover showed marked differences in cover and species composition across

the shelf (Figures 53-54). Mean % macroalgal cover was 14.7% across inshore sites, with a

minimum value of 1.7% at the 5-10 m depth zone of EME (in the understory below a huge %

gorgonian cover) and a highest value at the <5 m depth zone of OST with 32.7%. Mean %

macroalgal cover was 4.4% across mid-shelf sites, with a minimum value of 0.5% at the 10-15 m

depth zone of CON and a highest value at the 10-15 m zone of RES with 10.3%. Mean %

macroalgal cover was 18.5% across outer shelf sites, with a minimum value of 8.5% at the <5 m

depth zone of GAL and a highest value at the 10-15 m zone of GAL with 29.1%. Macroalgal

species composition varied from site to site (Figure 55). Brown and green macroalgae of

multiple mixed species were dominant across inshore sites. Brown macroalgae Dictyota spp.

were dominant across mid-shelf and outer shelf sites. Brown macroalgae Lobophora variegate

was also dominant, particularly at deeper outer shelf sites.


49

FIGURE 53. Cross-shelf spatial patterns of percent macroalgal cover.


50


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% M

acro

alg

ae

0

10

20

30

40

FIGURE 54. Percent macroalgal cover across sites. Red circles= Inshore sites; Green circles=



OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% M

acro

alg

ae

0

2

4

6

8

10

12

14

16

18

Green Macroalgae

C. racemosa

Red Macroalgae

A. taxiformis

Gracilaria sp.

R. textile

Brown macroalgae

Dictyopteris sp.

Dictyota sp.

L. variegata

Sargassum sp.

S. schroederi

FIGURE 55. Percent relative macroalgal cover across sites. Red circles= Inshore sites; Green


interval.


51

FIGURE 56. Cross-shelf spatial patterns of percent algal turf cover.


52


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% T

urf

0

5

10

15

20

25

30

35

40

FIGURE 57. Percent algal turf cover across sites. Red circles= Inshore sites; Green circles=


Percent algal turf cover showed also marked differences in cover across the shelf (Figures 56-

57). Mean % algal turf cover was 20.7% across inshore sites, with a minimum value of 10.0% at

the <5 m depth zone of EME and a highest value at the <5 m depth zone of RAT with 32.9%.

Mean % algal turf cover was 18.5% across mid-shelf sites, with a minimum value of 6.3% at the

10-15 m depth zone of RES and a highest value at the 10-15 m zone of RON with 27.9%. Mean

% algal turf cover was 16.3% across outer shelf sites, with a minimum value of 12.1% at the 10-

15 m depth zone of PPS and a highest value at the 5-10 m zone of NEG with 20.5%.

Percent Halimeda spp. cover showed also high variation among sites (Figures 58-59). Mean %

Halimeda spp. cover was 5.5% across inshore sites, with a minimum value of 0.04% at the <5 m

depth zone of EME and a highest value at the <5 m depth zone of GUA with 8.9%. Mean %

Halimeda spp. cover was 10.0% across mid-shelf sites, with a minimum value of 0.5% at the 10-

15 m depth zone of RON and a highest value at the <5 m zone of CON with 19.6%. Mean %

Halimeda spp. cover was 3.0% across outer shelf sites, with a minimum value below 0.2% at the

15-20 m depth zone of PPS and a highest value at the 5-10 m zone of NEG with 7.1%.


53

FIGURE 58. Cross-shelf spatial patterns of percent Halimeda spp. cover.


54


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% H

alim

ed

a

0

5

10

15

20

25

FIGURE 59. Percent Halimeda spp. cover across sites. Red circles= Inshore sites; Green


interval.

Percent crustose coralline algae (CCA) cover increased with distance from LBSP (Figures 60-

61). Mean % CCA cover was 0.7% across inshore sites, with a highest value of only 1.6% at the

<5 m depth zone of GUA and absent at the <5 m depth zone of both, EME and RAT. Mean %

CCA cover was 1.6% across mid-shelf sites, with a minimum value of 0.1% at the 10-15 m depth

zone of CON and a highest value at the <5 m zone of CON with 6.8%. Mean % CCA cover was

5.7% across outer shelf sites, with a minimum value below 1.1% at the 10-15 m depth zone of

PPS and a highest value at the <5 m zone of GAL with 15.7%. Porolithon pachydermum was the

most dominant CCA species across the entire shelf among at least five identified species (Figure

62). Erect calcareous algae (ECA), particularly Jania, showed a limited distribution.


55

FIGURE 60. Cross-shelf spatial patterns of percent crustose coralline algal (CCA) cover.


56


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% C

CA

0

5

10

15

20

25

FIGURE 61. Percent crustose coralline algal (CCA) cover across sites. Red circles= Inshore




OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% C

CA

+ E

CA

0

2

4

6

8

CCA

Hydrolithon sp.

M. mesomorphum

Peyssonnelia sp.

P. extensum

P. pachydermum

ECA

Jania sp.

L. ruptile

FIGURE 62. Percent relative crustose coralline (CCA) + erect calcareous algae (ECA) cover

across sites. Red circles= Inshore sites; Green circles= Mid-shelf sites; Black

circles= Outer shelf sites. Mean±95% confidence interval.


57

FIGURE 63. Cross-shelf spatial patterns of percent cyanobacterial cover.


58


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% C

ya

no

ba

cte

ria

0

10

20

30

40

FIGURE 64. Percent cyanobacterial cover across sites. Red circles= Inshore sites; Green


interval.

Cyanobacteria showed an interesting spatial pattern that reflected higher abundance across mid-

shelf sites, possibly influenced by effluents from either Boquerón Bay (RES), from Guanajibo

River (CON, RON, NEG), or from Mayagüez (PPS) (Figures 63-64). Mean % cyanobacterial

cover was 0.7% across inshore sites, with a highest value of only 2.2% at the <5 m depth zone of

OST and a lower value of less than 0.1% at the <5 m depth zone of GUA. Mean %

cyanobacterial cover was an impressive 17.1% across mid-shelf sites, with a minimum value of

2.2% at the 5-10 m depth zone of RES and a highest value at the 10-15 m zone of CON with

27.1%. Mean % cyanobacterial cover was 8.4% across outer shelf sites, with a minimum value

of 2.6% at the 5-10 m depth zone of GAL and a highest value at the 10-15 m zone of PPS with

16.9%, right at the shelf edge. These findings suggest that impacts of LBSP, particularly across

moderate to deeper reef zones are widespread across the entire shelf and appear to be largely

related to the complex cross-shelf oceanic circulation patterns.


59


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% C

ya

no

ba

cte

ria

0

2

4

6

8

10

12

14

Unk. cyanobacteria

Calothrix sp.

H. coccineum

Lyngbya sp.

Schizothrix sp.

Scytonema sp.

FIGURE 65. Percent relative cyanobacteria cover across sites. Red circles= Inshore sites;



Lyngbia spp. was the most common cyanobacterial taxa across the shelf, followed by

Hydrocoleum coccineum, particularly across mid- to outer shelf sites (Figure 65). Schizotrix spp.

was also common, particularly across mid-shelf sites. Calothrix spp. was particularly common at

selected mid-shelf sites RES and CON.

Percent frequency of recently dead corals (DCA) was generally low overall (Figure 66) and

averaged 0.4% across inshore sites, with a maximum of 1.8% at the <5 m depth zone of OST. No

DCA were documented at LAM. Mean % DCA was slightly less than 0.6% across mid-shelf

sites, with a highest value of 2.2% at the <5 m depth zone of RON. DCA were absent at the <5 m

and at the 10-15 m depth zones of RES. Mean % DCA was 0.8% across outer shelf sites, with

the lowest frequency at the 15-20 m depth zone on the vertical wall at PPS, and he highest at the

10-15 m zone of GAL with 2.8%.


60


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% D

CA

0

1

2

3

4

5

6

FIGURE 66. Percent frequency recently dead coral with algae (DCA) across sites. Red circles=



Percent frequency of diseased corals showed localized spatial variability probably influenced by

local water quality and LBSP (Figure 67) and averaged 3.1% across inshore sites, with a

maximum of 8.3% at the <5 m depth zone of OST, and a minimum prevalence of 0.4% at the <5

m depth zone of GUA. Mean % prevalence of diseased was 4.7% across mid-shelf sites, with a

an impressive highest prevalence of 28.0% at the 10-15 m depth zone of RES, mostly in the form

of black band disease (BBD) affecting Orbicella spp and Siderastrea siderea. The lowest mean

prevalence was observed across the <5 m depth zone of RES. Mean % prevalence of diseased

corals was 3.5% across outer shelf sites, with the lowest prevalence of 0.5% at the <0.5 m depth

zone at GAL, mostly in the form of patchy necrosis (PN) impacting exclusively Elkhorn coral

(Acropora palmata), and he highest at the 10-15 m zone of GAL with 2.8%, and the highest at

the 5-10 m depth zone of GAL, with 10.2%, mostly in the form of PN and in a minor degree

BBD. With the exception of mid-shelf 10-15 m depth zone of RES (11.9%), and <5 m depth

zone of OST (1.7%), coral bleaching was not a significant factor across sites (Figure 68).


61


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% D

ise

ase

d c

ora

ls

0

10

20

30

40

50

FIGURE 67. Percent prevalence of diseased corals across sites. Red circles= Inshore sites;




OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

% B

lea

ch

ed

0

5

10

15

20

25

FIGURE 68. Percent frequency diseased corals across sites. Red circles= Inshore sites; Green


interval.


62

FIGURE 69. Principal component ordination plot (PCO) of coral reef benthic community

structure as a function of water quality parameters and across a turbidity gradient

(bubbles). Clusters based on 60% similarity cutoff.

PCO analysis showed that benthic coral reef community structure across the western Puerto Rico

shelf showed three different clustering patterns as a function of water quality parameters (Figure

69). This PCO show results highlighting turbidity spatial patterns (bubbles). A first cluster of

inshore sites RAT, OST, LAM, and GUA were explained higher turbidity, PO4 and sea surface

temperature (SST), followed by salinity, conductivity. A second cluster was composed by EME

and RES, and was largely explained by NH4+, and in a lesser extent by chlorophyll-a and optical

brighteners concentration (OABs). A final cluster included mid-shelf CON, RON, and outer

shelf NEG, PPS, and GAL, which were largely explained by higher dissolved oxygen

-20 -10 0 10 20 30

PCO1 (34% of total variation)

-30

-20

-10

0

10

20

PC

O2

(1

8.6

% o

f to

tal va

ria

tio

n)

NTU

0.4

1.6

2.8

4

OSTLAM

GUA

RAT

EME

RES

CON

RON

NEG

PPS

GAL

TC

Sal

CondD.O

NTUChl-aOABs PO4

NH4+


63

concentration. This PCO explained 53% of the observed cross-shelf spatial variability in coral

recruit community structure.


structure as a function of water quality parameters and across a phosphate (PO4)

concentration (μM) gradient (bubbles). Clusters based on 60% similarity cutoff.

Figure 70 shows PO4 spatial patterns, while Figure 71 shows NH4+ spatial patterns. These three

examples suggest that these three parameters seem to affect coral reef benthic community

structure. A multivariate stepwise regression analysis using the BVSTEP routine confirmed that

the best combination of variables explaining benthic community structure spatial patterns was a

combination of turbidity, PO4 and NH4+ (correlation= 0.589, p<0.05).

-20 -10 0 10 20 30


-30

-20

-10

0

10

20

PC

O2

(1

8.6

% o

f to

tal va

ria

tio

n)

PO4

0.5

2

3.5

5

OSTLAM

GUA

RAT

EME

RES

CON

RON

NEG

PPS

GAL

TC

Sal

CondD.O

NTUChl-aOABs PO4

NH4+


64


structure as a function of water quality parameters and across a ionized

ammonium (NH4+) concentration (μM) gradient (bubbles). Clusters based on 60%

similarity cutoff.

Further, a multivariate correlation analysis using the RELATE routine and following a Spearman

Rank correlation procedure showed that there was a highly significant correlation between the

coral reef benthic community structure matrix and the matrix of water quality parameters (R=

0.42, p=0.0020).

A ‘linkage tree’ of coral reef benthic community structure based on the BIOENV routine to

environmental variables was carried out (Figure 72). Binary split on basis of the best single

environmental variable was thresholded to maximise the ANOSIM R statistic for the two groups

formed. The first split (A) in the assemblage data is between sites 8, 10, 11 (RON, PPS, GAL) in

-20 -10 0 10 20 30


-30

-20

-10

0

10

20P

CO

2 (

18

.6%

of

tota

l va

ria

tio

n)

NH4+

30

120

210

300

OSTLAM

GUA

RAT

EME

RES

CON

RON

NEG

PPS

GAL

TC

Sal

CondD.O

NTUChl-aOABs PO4

NH4+


65

the left side of the tree, and the second group was composed of sites 1-7, 9 (RAT, OST, LAM,

GUA, EME, RES, CON, NEG) in the right side of the tree. This division is also reflected in the

multi-dimensional scaling (MDS) plot of the test with an ANOSIM of R= 0.57 and B= 85.9%

(Figure 73). It is characterized by low ionized ammonium to the left of the tree (NH4+ Euclidean

distance < -0.677) at outer shelf sites PPS and GAL, and at mid-shelf site RON, in high NH4+

Euclidean distance > -0.546) to the right side of the tree (Figure 72). Alternatively, the same

split of sites was obtained by choosing high turbidity to the left of the tree (Turbidity Euclidean

distance < -0.555) at outer shelf sites PPS and GAL, and at mid-shelf site RON, and high

turbidity (Turbidity Euclidean distance < -0.677) to the right side of the tree (Figure 72).

ANOSIM R was the same whichever of the two variables was used as they gave the same split of

biotic data.

Moving down the tree, split (D) provides the second most strong explanation of spatial patterns

with a split between sites 1-6 (RAT, OST, LAM, GUA, EME, RES, CON) to the left and site 7

(NEG) to the right, with low optical brighteners to the left (OABs<0.304) and high optical

brighteners to the right (OABs >0.931), and an ANOSIM R= 1.00, and B= 80.7% (Figure 72).

But split A offered the best solution.

In synthesis, LINKTREE analysis showed that variation in NH4+ and turbidity explained most of

the spatial variation observed in coral reef benthic community structure. In conclusion, coral reef

benthic community structure is being largely influenced by LBSP stress gradients across the

entire western Puerto Rican shelf. Coral assemblages are showing signs of stress gradient and so

are compromising the long-term reef accretion sustainability and ecosystem resilience.


66

A->B,C= R: 0.57, B%: 85.9, NH4+<-0.677(>-0.546), NTU<-0.555(>-0.463) B->(11),(8-10)= R: 0.56, B%: 36.1, Sal<-1.84(>-0.791), Cond<-2.01(>-1.22), Chl-a<-0.81(>-0.319), D.O>1.72(<1.39) C->(5,6),D= R: 0.47 B%: 67.2, Sal<-0.267(>8.16E-2), D.O>-0.226(<-0.418) D->E,(7)= R: 1.00 B%: 80.7, OABs<0.304(>0.931) E->(1,2),(3,4)= R: 0.75 B%: 21.7, Chl-a>1.16(<6.2E-2), NH4

+<-9.43E-2(>0.756), TC<0.212(>1.03),

Sal>0.777(<0.256), PO4>0.842(<0.4), OABs>4.87E-2(<-8.29E-2), NTU>1.17(<1.06), D.O<-0.528(>-0.453) Sample numbers for plot 1)RAT;2)OST;3)LAM;4)GUA;5)EME;6)RES;7)CON;8)RON;9)NEG;10) PPS;11)GAL

FIGURE 72. LINKTREE analysis of coral reef benthic community structure as a function of

water quality parameters. Data was log10-transformed and normalized for

analysis. Numbers are Euclidean distances. B%= ‘between group’ average rank as

% of maximum possible.

0

20

40

60

80

100B

%

11 8-10B

1,2 3,4E

7D

5,6C

A


67

FIGURE 73. Multi-dimensional scaling (MDS) plot of the first stage in a ‘linkage tree’ of coral

reef benthic community structure to environmental variables. Binary split on basis

of the best single environmental variable, thresholded to maximise the ANOSIM

R statistic for the two groups formed.

So far, this study has confirmed that benthic coral reef communities are showing evidence of

impacts by chronic LBSP in the form of altered spatial distributions of multiple coral and algal

taxa, with particular distance gradients (i.e., increasing percent living coral cover with increasing

distance from LBSP). Multivariate PCO further confirmed these patterns. In spite of the fact that

multiple coral reef show high benthic spatial heterogeneity due to structural construction by

scleractian corals, percent living coral cover remains low, regardless of depth and distance from

the shore. This suggests that coral reef benthic communities are showing a highly limited natural

recovery ability, probably as a combined result of chronic LBSP impacts (which foster increased

algal growth) and impacts by climate change-related factors (see Hernández-Delgado et al.,


68

2014). Therefore, coral recruit spatial patterns have become a critical component to address the

spatial patterns of LBSP influences across the shelf.

TABLE 3. Summary of a three-way PERMANOVA test of coral recruit community structure


F p

Geographic location 2,147 10.31 <0.0001 Site 10,139 4.39 <0.0001 Depth 3,146 3.10 0.0004 Location x Site 10,139 4.39 <0.0001 Location x Depth 8,141 3.97 <0.0001 Site x Depth 22,127 3.01 <0.0001 Location x Site x Depth 22,127 3.01 <0.0001

C. Coral recruit communities

Coral recruit community structure showed a highly significant variation among all geographic

zones, sites and depth zones, suggesting a strong effect associated to the LBSP gradient and to

depth (Table 3). PCO analysis suggested that coral recruit community structure produced three

general clustering patterns (Figure 74). A first cluster was dominated by inshore sites EME and

RAT, which had very limited abundance of Scleractinians. This cluster was explained by

Gorgonia ventalina recruits. A second cluster was composed by inshore sites OST and GUA,

which was largely explained by Siderastrea radians recruits, which is an ephemeral species,

dominant under high disturbance regimes, characterized by frequent (lunar) recruitment cycles,

but very high mortality rates. The final large cluster is composed by mid-shelf and outer shelf

sites plus inshore site LAM, and is explained by a combination of several Scleractinian species.

Inshore site LAM was explained by S. radians, and RES by G. ventalina. RON was largely

explained by Millepora alcicornis recruitment. PPS was better explained by S. siderea recruits,

while NEG, CON, and GAL were better explained by a group of Scleractinian recruits, including

Porites porites, Undaria tenuifolia, U. agaricites, and Agaricia lamarcki. Recruits of these

species were largely absent from inshore sites.


69

FIGURE 74. Principal component ordination plot of coral recruit communities among study

sites. Clusters based on 40% similarity cut off level. Vectors based on a 0.70

correlation. This model explained 60.9% of the observed variation.

A spatial gradient of increasing coral recruit abundance with increasing distance from LBSP was

also evident for several species, including Porites porites, Undaria agaricites, Siderastrea

siderea, and Orbicella faveolata (Figure 75). Coral recruit density averaged 1.1 colonies/m2

across inshore reef sites, with the lowest value of 0.4 colonies/m2 at the <5 m depth zone of RAT

and the highest value of 2.1 colonies/m2 at the <5 m depth zone of LAM (Figure 76). Coral

recruit density averaged 4.1 colonies/m2 across mid-shelf sites, with the lowest value of 1.0

colonies/m2 at the 10-15 m depth zone at RES and the highest value of 8.4 colonies/m

2 at the <5

m depth zone of RON. Mean coral recruit density across outer shelf sites was 7.7 colonies/m2

across mid-shelf sites, with the lowest value of 4.8 colonies/m2 at the <5 m depth zone at NEG

and the highest value of 16.0 colonies/m2 at the 10-15 m depth zone of GAL.

-60 -40 -20 0 20 40 60


-50

0

50P

CO

2 (

21

.3%

of

tota

l va

ria

tio

n)

Similarity40

RAT

OST

LAM

GUA

EME

RES

CON

RON

NEG

PPSGAL

Gven

Past

PporUagaUten Alam

Srad

Ssid

Malc


70

FIGURE 75. Bubble plots based on a MDS analysis of the spatial distribution of four Scleractinian coral species along a LBSP

gradient. From top left: A) Porites porites (Ppor); B) Undaria agaricites (Uaga); C) Siderastrea siderea (Ssid); and D)

Orbicella faveolata. Clustering patterns based on 40% similarity cut off level.

Ppor

7E-2

0.28

0.49

0.7

RAT

OST

LAM

GUA

EMERES

CON

RON

NEG

PPSGAL

2D Stress: 0.11 Uaga

0.2

0.8

1.4

2

RAT

OST

LAM

GUA

EMERES

CON

RON

NEG

PPSGAL

2D Stress: 0.11

Ssid

0.2

0.8

1.4

2

RAT

OST

LAM

GUA

EMERES

CON

RON

NEG

PPSGAL

2D Stress: 0.11 Ofav

9E-2

0.36

0.63

0.9

RAT

OST

LAM

GUA

EMERES

CON

RON

NEG

PPSGAL

2D Stress: 0.11


71

Western PR shelf


OS

T-I

LA

M-I

GU

A-I

RA

T-I

EM

E-I

EM

E-I

I

RE

S-I

RE

S-I

I

RE

S-I

II

CO

N-I

CO

N-I

I

CO

N-I

II

RO

N-I

RO

N-I

I

RO

N-I

II

NE

G-I

NE

G-I

I

NE

G-I

II

PP

S-I

II

PP

S-I

V

GA

L-I

GA

L-I

I

GA

L-I

II

Co

ral re

cru

it d

en

sity (

#/m

2)

0.1

1

10

100

FIGURE 76. Total coral recruit density across sites. Red bars= Inshore sites; Gold bars= Mid-

shelf sites; Black bars= Outer shelf sites. Mean±95% confidence interval.

f = 11.13*exp(-0.88*x)

r2=0.5520, p=0.0088

NTU

0 1 2 3 4

Cora

l re

cru

it d

ensity (

#/m

2)

0

5

10

15

20

25

Recruit density

Regresssion line

95% Confidence Band

FIGURE 77. Non-linear regression between total coral recruit density and turbidity.


72

There was a significant (r2=0.5520, p=0.0088) non-linear negative correlation between coral

recruit density and water turbidity (Figure 77). There was also a significant (r2=0.5203,

p=0.0122) linear positive correlation between coral recruit density and dissolved oxygen

concentration (Figure 78). Coral recruit density showed also a significant (r2=0.7464, p=0.0006)

non-linear negative correlation between and NH4+ (Figure 79).

Scleractinian coral recruit species distribution was largely variable among sites across the LBSP

distance gradient (Figure 80). Porites astreoides, Siderastrea siderea, and S. radians were the

dominant Scleractinian recruit species across inshore sites. Siderastrea siderea, P. astreoides,

and P. porites were the dominant Scleractinian recruits across mid-shelf sites. Siderastrea

siderea, P. astreoides, and Undaria agaricites were the dominant Scleractinian recruits across

outer shelf sites.

Octocoral recruit density was low across the entire shelf, with mean densities that never

exceeded 0.2 colonies/m2 across inshore sites, 1.2 colonies/m

2 across mid-shelf sites, and 0.8

colonies/m2 across outer shelf sites (Figure 81). Octocoral recruits species composition showed

a significant difference among all geographic locations. Gorgonia ventalina recruits were present

only at the <5 m depth zone of EME, while Plexaura hommomalla was present only at the <5 m

depth zone of LAM. No other octocoral recruits were documented across other inshore sites.

Briareum asbestinum was present only at the 5-10 m depth zone of RES, while Eunicea

calyculata was present at the <5 m depth zone of CON. No other octocoral recruits were present

across mid-shelf sites. Muriceopsis flavida and E. fusca were present at the <5 m depth zone of

GAL, while P. hommomalla was only present at the 5-10 m depth zone of GAL.


73

f = -11.02+3.3*x

r2=0.5203, p=0.0122

Dissolved oxygen (mg/L)

3.5 4.0 4.5 5.0 5.5 6.0

Cora

l re

cru

it d

ensity (

#/m

2)

0

2

4

6

8

10

12

Recruit density

Regression line

95% Confidence Band

FIGURE 78. Linear regression between total coral recruit density and dissolved oxygen

concentration.

f = 11.13*exp(-0.88*x)

r2=0.7464, p=0.0006

NH4+ (uM)

0 50 100 150 200 250 300

Cora

l re

cru

it d

ensity (

#/m

2)

0

2

4

6

8

10

12

14 Recruit density

Regression line

95% Confidence Band

FIGURE 79. Non-linear regression between total coral recruit density and. ammonium (NH4+)

concentration.


74

A

Inshore sites

OST-I LAM-1 GUA-I RAT-I EME-I EME-II

Scle

ractin

ian

s +

Hyd

roco

rals

(#

/m2)

0

2

4

6

8

10

12

14

16

18

M. alcicornis

P. porites

P. divaricata

P. furcata

P. astreoides

S. radians

S. siderea

U. agaricites

U. tenuifolia

U. humilis

A. lamarcki

C. natans

P. strigosa

D. labyrynthiformis

M. cavernosa

O. annularis

M. meandrites

M. areolata

S. bournoni

B

Midshelf sites

RES-I RES-II RES-III CON-I CON-II CON-III RON-I RON-II RON-III

Scle

ractin

ian

s +

Hyd

roco

rals

(#

/m2)

0

2

4

6

8

10

12

14

16

18

C

Outer shelf sites

NEG-I NEG-II NEG-III PPS-III PPS-IV GAL-I GAL-II GAL-III

Scle

ractin

ians +

Hyd

roco

rals

(#

/m2)

0

2

4

6

8

10

12

14

16

18

FIGURE 80. Mean scleractinian coral and hydrocoral recruit density across study sites: A)

Inshore reefs; B) Mid-shelf reefs; and C) Outer shelf reefs. I= <5 m, II= 5-10 m,

III= 10-15 m; IV= 15-20 m.


75

A

Inshore sites

OST-I LAM-1 GUA-I RAT-I EME-I EME-II

Octo

co

rals

(#

/m2)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

G. ventalina

B. asbestinum

E. fusca

E. calyculata

M. flavida

P. hommomalla

B

Midshelf sites


Octo

co

rals

(#

/m2)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

C

Outer shelf sites


Octo

co

rals

(#

/m2)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

FIGURE 81. Mean octocoral recruit density across study sites: A) Inshore reefs; B) Mid-shelf

reefs; and C) Outer shelf reefs. I= <5 m, II= 5-10 m, III= 10-15 m; IV= 15-20 m.


76

FIGURE 82. Principal component ordination plot (PCO) of coral recruit community structure

as a function of water quality parameters and across a turbidity gradient (bubbles).

Clusters based on 40% similarity cutoff level. This model explained 60.9% of the

observed variation.

PCO analysis showed that coral recruit community structure across the western Puerto Rico shelf

showed three different clustering patterns as a function of water quality parameters (Figure 82).

This PCO show results highlighting turbidity spatial patterns (bubbles). A first cluster of inshore

sites RAT and EME were explained higher turbidity, PO4 and NH4+. A second cluster was

composed by OST and GUA, and was largely explained by salinity. A final large cluster was

explained by the rest of the water quality parameters, with RES explained by OABs and

chlorophyll-a, and the rest of the mid- and outer shelf sites explained by dissolved oxygen

concentration. This PCO explained 61% of the observed cross-shelf spatial variability in coral

recruit community structure.

-60 -40 -20 0 20 40 60


-50

0

50P

CO

2 (

21

.3%

of

tota

l va

ria

tio

n)

NTU

-1.6

-0.4

0.8

2

RAT

OST

LAM

GUA

EME

RES

CON

RON

NEG

PPSGAL

Sal

Cond

D.O

NTU

Chl-a

OABs

PO4 (uM)

NH4+ (uM)


77


as a function of water quality parameters and across a phosphate (PO4)

concentration (μM) gradient (bubbles). Clusters based on 40% similarity cutoff

level. This model explained 60.9% of the observed variation.

Figure 83 shows PO4 spatial patterns, while Figure 84 shows NH4+ spatial patterns. These three

examples suggest that these three parameters seem to affect coral recruit community structure.

BIOENV and BVSTEP analyses showed that the coral recruit biotic matrix and the water quality

matrix showed a correlation of 0.463 based on the combination of turbidity, chlorophyll-a, and

PO4 concentrations. This suggests that although LBSP could have been an important factor

influencing coral recruit spatial patterns, other potential non-local, regional-global factors (i.e.,

climate change) could have significantly influenced coral recruitment trends across surveyed

sites.

-60 -40 -20 0 20 40 60


-50

0

50P

CO

2 (

21

.3%

of

tota

l va

ria

tio

n)

PO4 (uM)

-1.6

-0.4

0.8

2

RAT

OST

LAM

GUA

EME

RES

CON

RON

NEG

PPSGAL

Sal

Cond

D.O

NTU

Chl-a

OABs

PO4 (uM)

NH4+ (uM)


78


as a function of water quality parameters and across a ionized ammonium (NH4+)

concentration (μM) gradient (bubbles). Clusters based on 60% similarity cutoff

level. This model explained 60.9% of the observed variation.

A ‘linkage tree’ of coral recruit community structure based on the BIOENV routine to



formed. The first split (A) in the assemblage data is between site 1 (RAT) in the left side of the

tree, and the rest in the right side of the tree. This division is also reflected in the MDS plot of the

test with an ANOSIM of R= 0.68 and B= 85.9% (Figure 86). It is characterized by a high

turbidity (Euclidean distance > 1.45) to the left of the tree at inshore site RAT, and low turbidity

(Euclidean distance < 1.17) to the right side of the graph (Figure 85).

-60 -40 -20 0 20 40 60


-50

0

50P

CO

2 (

21

.3%

of

tota

l va

ria

tio

n)

NH4+ (uM)

-0.7

0.2

1.1

2

RAT

OST

LAM

GUA

EME

RES

CON

RON

NEG

PPSGAL

Sal

Cond

D.O

NTU

Chl-a

OABs

PO4 (uM)

NH4+ (uM)


79

A->(1),B= R: 0.68, B%: 85.9, NTU>1.45(<1.17), D.O<-1.03(>-1.03) B->C,(4,5)= R: 0.67, B%: 69.9, NH4+(<1.09(>1.12) C->(2),D= R: 0.78 B%: 57.4, Chl-a>1.46(<0.286), PO4 (uM)>0.888(<0.453), Cond>0.898(<0.755), NTU>1.17(<1.06) D->E,(11)= R: 0.44 B%: 36.6, Cond>-1.22(<-2.01), Sal>-1.32(<-1.84), D.O<1.39(>1.72) E->(3,6,10),(7-9)= R: 0.67 B%: 30.4, Sal<8.16E-2(>0.256) Sample numbers for plot 1)RAT;2)OST;3)LAM;4)GUA;5)EME;6)RES;7)CON;8)RON;9)NEG;10) PPS;11)GAL

FIGURE 85. LINKTREE analysis of coral recruit community structure as a function of water

quality parameters. Data was log10-transformed and normalized for analysis.

Numbers are Euclidean distances. B%= ‘between group’ average rank as % of

maximum possible.

0

20

40

60

80

100B

%

3,6,10 7-9E

11D

2C

4,5B

1A


80

Alternatively, the same split of sites was obtained with low dissolved oxygen concentration

(Euclidean distance < -1.03) to the left of the tree at inshore site RAT, and high dissolved oxygen

concentration (Euclidean distance > -1.03) to the right of the tree (Figure 85). ANOSIM R was

the same whichever of the two variables was used as they gave the same split of biotic data.

Moving down the tree, split (B) provides the second most strong explanation of spatial patterns

with a split between sites 4-5 (GUA, EME) to the right and the rest to the left, with low NH4+ to

the left (Euclidean distance < 1.09) and high NH4+ to the right at GUA and EME (Euclidean

distance > 1.12), and an ANOSIM R= 0.67, and B= 69.9% (Figure 87). Split A offered the best

solution.

In synthesis, LINKTREE analysis showed that variation in turbidity and dissolved oxygen

concentration explained most of the spatial variation observed in coral recruit community

structure. In conclusion, coral recruit community structure is also being largely influenced by

LBSP stress gradients across the entire western Puerto Rican shelf. Coral recruit assemblages are

also showing signs of a stress gradient and so are compromising the long-term reef recovery

ability, accretion sustainability and ecosystem resilience across mot reefs.


81


recruit community structure to environmental variables. Binary split on basis of

the best single environmental variable, thresholded to maximise the ANOSIM R

statistic for the two groups formed.


82

FIGURE 87. Multi-dimensional scaling (MDS) plot of the second stage in a ‘linkage tree’ of

coral recruit community structure to environmental variables. Binary split on

basis of the best single environmental variable, thresholded to maximise the

ANOSIM R statistic for the two groups formed.


83

TABLE 4. Summary of a three-way PERMANOVA test of reef fish community structure.


F p

Geographic location 2,94 4.94 <0.0001 Site 10,86 3.59 <0.0001 Depth 3,93 2.70 <0.0001 Location x Site 11,85 3.45 <0.0001 Location x Depth 9,87 2.75 <0.0001 Site x Depth 23,73 2.67 <0.0001 Location x Site x Depth 24,72 2.56 <0.0001

D. Coral reef fish communities

Coral reef fish community structure showed a highly significant variation among all geographic

zones, sites and depth zones, suggesting a strong effect associated to geographic location, to the

LBSP gradient and to depth (Table 4). PCO analysis suggested that coral reef fish community

structure produced four general clustering patterns (Figure 88). A first cluster was composed by

inshore sites RAT and OST, where RAT was explained by the Princess parrotfish (Scarus

taeniopterus) and the Beaugregory (Stegastes leucostictus), and where OST was explained by

Yellowtail parrotfish (Sparisoma rubripinne). There was also a second large cluster composed

by inshore sites LAM, GUA, and EME, mid-shelf sites RES, CON, and RON, and outer shelf

site GAL. These were explained by the Bluehead wrasse (Thalassoma bifasciatum) and by the

Sharpnose puffer (Canthigaster rostrata). Outer shelf site NEG constituted an independent

cluster mostly explained by multiple fish species, including the French angelfish (Pomacanthus

paru), the Creole wrasse (Clepticus parrae), and the Brown chromis (Chromis multilineata). But

also a consortium of five fish species explained the separation of NEG as an individual cluster,

including the Tobacco fish (Serranus tabacarius), the Bermuda chub (Kyphosus sectatrix), the

Atlantic spadefish (Chaetodipterus faber), the Rainbow parrotfish (Scarus guacamaia), and the

Frillfin goby (Bathygobius soporator). Outer shelf site PPS also constituted an independent

cluster explained by the Yellowtail snapper, Ocyurus chrysurus, and by the Cleaning goby,

Elacatinus genie, and by wrasse Bodianus pulchellus. These spatial patterns evidenced an

unequivocal cross-shelf gradient of fish species distribution influenced by reef position and

condition, depth, and water quality.


84

FIGURE 88. Principal component ordination plot of coral reef fish communities among study

sites. Clusters based on 50% similarity cut off level. Vectors based on a 0.70

correlation. This model explained 60.9% of the observed variation.

PCO analysis also revealed that the observed spatial clustering pattern in fish community

structure was influenced by water quality spatial patterns (Figure 89). The inshore cluster

composed by RAT and OST was largely explained by turbidity, chlorophyll-a, and salinity in a

lesser extent. The large cluster formed by inshore sites LAM, GUA, and EME, the mid-shelf

RES, CON, and RON, and the outer shelf GAL were explained by PO4 and NH4+ concentrations,

and in a lesser extent by conductivity and OABs, particularly across inshore sites. Outer shelf

sites PPS and NEG were most likely explained by dissolved oxygen concentration.

-40 -20 0 20 40


-40

-20

0

20

40P

CO

2 (

19

.9%

of

tota

l va

ria

tio

n)

Similarity50

OST

LAMGUA

RAT

EME

RES

CONRON

NEG

PPS

GAL

Stab

Ochr

KsecCfabPpar Sleu

Cmul

Bpul

Tbif

Cpar

Srub

StaeSguaBsop

Egen

Crst


85

FIGURE 89. Principal component ordination plot of coral reef fish communities among study

sites based on water quality spatial patterns. Clusters based on 50% similarity cut

off level. This model explained 60.9% of the observed variation.

A % cumulative dominance plot of fish species also evidenced the observed cross-shelf spatial

gradient of fish species richness (Figure 90). A SIMPER test of indicator fish species based on

biomass estimates revealed that parrotfishes (Scaridae) were the most significant indicators of

spatial patterns across most surveyed sites (Table 5). The Striped parrotfish (Scarus iserti) was

the most common indicator species across 6 out of 11 sites (55%). Four other parrotfish species,

including Sparisoma rubripinne, Sp. viride, Sp. aurofrenatum, and Scarus taeniopterus, as well

as Labrid Clepticus parrae were also indicator species at individual sites. These were the most

dominant (=higher mean biomass) across sites.

-40 -20 0 20 40


-40

-20

0

20

40P

CO

2 (

19.9

% o

f to

tal va

riation)

Similarity50

OST

LAMGUA

RAT

EME

RES

CONRON

NEG

PPS

GAL

Sal

Cond

D.O

NTU

Chl-a

OABs

PO4 (uM)

NH4+ (uM)


86

FIGURE 90. Cumulative fish species dominance plot.

1 10 100

Species rank

0

20

40

60

80

100C

um

ula

tive

Do

min

an

ce

%OST

LAM

GUA

RAT

EME

RES

CON

RON

NEG

PPS

GAL


87

TABLE 5. Summary of SIMPER test of the three most dominant coral reef fish species

across sites.

Species/Site % Contrib. Cum. % OST Sparisoma rubripinne Acanthurus coeruleus Scarus iserti

31.73 16.40 2.06

31.73 48.13 63.36

LAM Scarus iserti Sparisoma viride Sparisoma rubripinne

33.50 23.24 8.07

33.50 56.74 64.81

GUA Scarus iserti Acanthurus bahianus Sparisoma viride

29.99 10.85 10.24

29.99 40.85 51.09

RAT Scarus taeniopterus Stegastes planifrons Scarus iserti

26.66 21.96 12.49

26.66 48.62 61.11

EME Scarus iserti Acanthurus coeruleus Acanthurus bahianus

23.67 21.08 15.46

23.67 44.75 60.21

RES Scarus iserti Sparisoma viride Sparisoma aurofrenatum

27.57 11.80 11.74

27.57 39.37 51.10

CON Sparisoma viride Sparisoma aurofrenatum Scarus iserti

18.37 15.46 14.41

18.37 33.83 48.24

RON Sparisoma aurofrenatum Sparisoma viride Scarus iserti

20.37 14.46 13.65

20.37 34.83 48.48

NEG Clepticus parrae Sparisoma viride Scarus iserti

29.84 14.02 9.74

29.84 43.86 53.61

PPS Scarus iserti Stegastes partitus Cephalopholis fulva

20.84 9.97 8.84

20.84 30.82 39.65

GAL Scarus iserti Melichthys niger Sparisoma aurofrenatum

15.24 12.07 9.19

15.24 27.30 36.50


88

TABLE 6. Summary of percent dissimilarities in coral reef fish assemblages across sites.

Sites LAM GUA RAT EME RES CON RON NEG PPS GAL

OST 74 73 66 73 75 71 78 86 82 74

LAM - 72 73 75 67 69 73 81 80 72

GUA - 76 71 69 70 75 84 80 74

RAT - 75 74 69 76 80 81 71

EME - 66 69 75 82 76 74

RES - 63 66 77 74 66

CON - 62 76 74 62

RON - 77 75 66

NEG - 79 76

PPS - 77

GAL -

There was also a spatial gradient of increasing mean dissimilarities in fish assemblages with

increasing distance from the shore (Table 6). Mean % dissimilarity within inshore sites was

72.8%, with a range of 66 to 76%. Mean % dissimilarity within mid-shelf sites was 63.7%, with

a range of 62 to 66%. Mean % dissimilarity within outer-shelf sites was 77.3%, with a range of

76 to 79%. Mean % dissimilarity between inshore and mid-shelf sites was 71.7%, and between

inshore and outer shelf sites was 78.5%. Mean % dissimilarity between mid-shelf and outer shelf

sites was 71.9%. This suggests that geographic distance, and potentially reef condition and a

water quality stress gradient could have influenced cross-shelf fish diversity distribution.


89

TABLE 7. Indicator coral reef fish species analyses across sites based on SIMPER test.

Sites LAM GUA RAT EME RES CON RON NEG PPS GAL

OST Sise˄ Srub˂ Stae˂ Srub˂ Srub˂ Srub˂ Srub˂ Cpar˄ Srub˂ Srub˂

LAM - Sise˂ Sise˂ Sise˂ Sise< Sise˂ Sise˂ Cpar˄ Sise˂ Sise˂

GUA - Stae˂ Svir˂ Svir˄ Svir˄ Saur˄ Cpar˄ Cpar˄ Mnig˄

RAT - Stae˂ Stae˂ Stae˂ Stae˂ Cpar˄ Cpar˄ Mnig˄

EME - Sise˄ Saur˄ Sbar˄ Cpar˄ Cpar˄ Mnig˄

RES - Sise˂ Sbar˄ Cpar˄ Cpar˄ Mnig˄

CON - Sbar˄ Cpar˄ Cpar˄ Mnig˄

RON - Cpar˄ Cpar˄ Mnig˄

NEG - Cpar˂ Cpar˂

PPS - Crub˂

GAL -

˄= Higher biomass at the site above. ˂= Higher biomass at the site in the left.

SIMPER analysis also allowed determining indicator species of spatial differences among all

sites (Table 7). The Striped parrotfish (Scarus iserti), followed by the Princess parrotfish (Sc.

taeniopterus), was the most important indicator species of differences among inshore sites. The

Barracuda (Sphyraena barracuda) was the most important indicator species of differences

among mid-shelf sites, while the Creole wrasse (Clepticus parrae) was the most important

indicator among outer shelf sites. Differences between inshore and mid-shelf sites were mostly

explained by three parrotfish species, Sparisoma rubripinne, Sc. taeniopterus, and Sc. iserti.

Differences between inshore sites and outer shelf sites were mostly explained by C. parrae,

followed by the Black durgon (Melichthys niger). Differences between mid-shelf and outer shelf

sites were also mostly explained by these two species, and in a lower extent by the Bar jack

(Carangoides ruber).


90

A->(9,10),B= R: 0.83, B%: 93.2, PO4 < -1.57 (> -0.46), OABs < -1.33 (> -1.19) B->C,(1,4)= R: 0.90, B%: 68.4, Chl-a < 0.88 (>1.16),NTU < 1.14 (> 1.17) C->(7,8,11),D= R: 0.67 B%: 36.0, NH4

+ < -0.515 (>0.76)

D->(2,3,6), (5)= R: 0.78 B%: 25.8, OABs < -8.29E-2 (> 1.41), PO4 < 0.4 (> 1.34), Chl-a <6.2E-2 (> 0.88), NH4

+ <1.12 (> 1.77), NTU <1.06 (> 1.14)

Sample numbers for plot 1)RAT;2)OST;3)LAM;4)GUA;5)EME;6)RES;7)CON;8)RON;9)NEG;10) PPS;11)GAL

FIGURE 91. LINKTREE analysis of coral reef fish community structure as a function of water

quality parameters. Data was log10-transformed and normalized for analysis.

Numbers are Euclidean distances. B%= ‘between group’ average rank as % of

maximum possible.

A ‘linkage tree’ of coral reef fish community structure based on the BIOENV routine to



formed. The first split (A) in the assemblage data is between sites 9 (NEG) and 10 (PPS) in the

left side of the tree, and the rest in the right side of the tree. This division is also reflected in the

0

20

40

60

80

100B

%

2,3,6 5D

7,8,11C

1,4B

9,10A


91

MDS plot of the test with an ANOSIM of R= 0.83 and B= 93.2% (Figure 92). It is characterized

by a low PO4 concentration (Euclidean distance > 1.45) to the left of the tree at outer shelf sites

NEG and PPS, and a high PO4 concentration (Euclidean distance < 1.17) to the right side of the

graph (Figure 91). Alternatively, the same split of sites was obtained with low OABs

concentration (Euclidean distance < -1.33) to the left of the tree at outer shelf sites NEG and

PPS, and high OABs concentration (Euclidean distance > -1.19) to the right of the tree (Figure

91). ANOSIM R was the same whichever of the two variables was used as they gave the same

split of biotic data.

Moving down the tree, split (B) provides the second most strong explanation of spatial patterns

with a split between inshore sites 1 (RAT) and 5 (GUA) to the right and the rest to the left, with

low chl-a concentration to the left (Euclidean distance < 0.88) and high to the right at RAT and

GUA (Euclidean distance > 1.16), and an ANOSIM R= 0.90, and B= 68.4% (Figure 93).

Alternatively, the same split of sites was obtained with low turbidity (Euclidean distance < 1.14)

to the left of the tree, and high turbidity (Euclidean distance > 1.17) to the right of the tree at

RAT and GUA (Figure 91). ANOSIM R was the same whichever of the two variables was used

as they gave the same split of biotic data. But split A offered the best solution.

LINKTREE analysis showed that variation in PO4 and OABs concentrations explained most of

the spatial variation observed in the cross-shelf coral reef fish community structure. A

multivariate correlation using the RELATE routine showed a strong significant correlation

(Rho=0.362, p=0.0027) between the entire fish community matrix and the water quality matrix.

Further, BIOENV and BVSTEP analyses showed that the coral reef fish community matrix and

the water quality matrix showed a correlation of 0.626 (p<0.05) based on PO4 concentration.

Other factors, besides LBSP, could have also influenced fish community spatial patterns,

including fishing impacts and habitat decline associated to climate change.


92


reef fish community structure to environmental variables. Binary split on basis of

the best single environmental variable, thresholded to maximise the ANOSIM R

statistic for the two groups formed.


93

FIGURE 93. Multi-dimensional scaling (MDS) plot of the second stage in a ‘linkage tree’ of

coral reef fish community structure to environmental variables. Binary split on

basis of the best single environmental variable, thresholded to maximise the

ANOSIM R statistic for the two groups formed.

Mean fish species richness increased with increasing distance from the shore along the

documented LBSP gradient (Figure 94). Species richness averaged 12.6/count across inshore

sites, with a minimum of 9.3 species/count at the 10-15 m depth zone of EME and a maximum

of 14.2 species/count at the <5 m depth zone of GUA. Species richness increased to 18.8/count

across mid-shelf sites, with a minimum 14.5/count at the <5 m depth zone of RES and a

maximum of 27/count at the 10-15 m depth zone of RON. Species richness also averaged

18.8/count across outer shelf sites, with a minimum of 14/count at the 5-10 m depth zone of

NEG and at the 10-15 m depth zone of PPS, and a maximum of 27/count at the 10-15 m depth

zone of GAL. Significantly polluted and degraded EME site showed the lowest species richness.


94

A

Inshore sites

OST-I LAM-I GUA-I RAT-I EME-I EME-II EME-III

Sp

ecie

s r

ich

ne

ss

0

5

10

15

20

25

30

35

B

Midshelf sites


Sp

ecie

s r

ich

ne

ss

0

5

10

15

20

25

30

35

C

Outer shelf sites


Sp

ecie

s r

ich

ne

ss

0

5

10

15

20

25

30

35

FIGURE 94. Fish species richness across study sites: A) Inshore reefs; B) Mid-shelf reefs; and

C) Outer shelf reefs. Mean±95% confidence interval. I= <5 m, II= 5-10 m, III=

10-15 m; IV= 15-20 m.


95

Mean fish abundance showed a similar increase with increasing distance from the shore along

the documented LBSP gradient (Figure 95). Fish abundance averaged 92.2/count across inshore

sites, with a minimum of 23 individuals/count at the 10-15 m depth zone of EME and a

maximum of 204 individuals/count at the <5 m depth zone of LAM, mostly juvenile and semi-

adult parrotfishes (Scaridae). Abundance increased to 166/count across mid-shelf sites, with a

minimum of 87/count at the 10-15 m depth zone of RES and a maximum of 307/count at the 10-

15 m depth zone of RON. Abundance further increased to 409/count across outer shelf sites, with

a minimum of 98/count at the 5-10 m depth zone of GAL and a maximum of 1,336/count at the

10-15 m depth zone of NEG. The lowest fish abundances were documented across the most

degraded and polluted reef sites.

Mean fish species diversity (H’n) followed the classical Connell (1978) intermediate disturbance

hypothesis curve with higher H’n under intermediate disturbance levels and lower H’n under

either low or frequent disturbance levels. H’n averaged 1.9567 across inshore sites, with a

minimum of 1.7631 at the 10-15 m depth zone of EME and a maximum of 2.2062 at the 5-10 m

depth zone of EME (Figure 96). H’n increased to 2.1301 across mid-shelf sites, with a minimum

of 1.8693 at the 5-10 m depth zone of RON and a maximum of 2.3401 at the 10-15 m depth zone

of RES. H’n averaged 1.8089 across outer shelf sites, with a minimum of 1.0080 at the 10-15 m

depth zone of NEG and a maximum of 2.3195 at the 10-15 m depth zone of GAL. Declining H’n

was often the result of dominance by juvenile and semi-adult parrotfish (Scaridae) guilds across

shallow inshore sites, and by Creole wrasse (Clepticus parrae) and Yellowtail snapper (Ocyurus

chrysurus) schools across outer shelf sites. However, benthic structural complexity, sediment

deposition and the condition of benthic habitats did influence these results.

Mean fish species evenness (J’n) showed a declining trend with increased distance from the

shore and increased dominance by selected species across outer shelf sites (Figure 97). J’n

averaged 0.7877 across inshore sites, with a minimum of 0.6865 at the <5 m depth zone of OST

and a maximum of 0.8588 at the 10-15 m depth zone of EME. J’n increased to 0.7371 across

mid-shelf sites, with a minimum of 0.6221 at the 10-15 m depth zone of RON and a maximum of

0.8230 at the <5 m depth zone of RON. J’n averaged 0.6162 across outer shelf sites, with a

minimum of 0.3397 at the 10-15 m depth zone of NEG and a maximum of 0.7613 at the 5-10 m

depth zone of GAL. Declining J’n was the result of dominance by C. parrae) and O. chrysurus.


96

A

Inshore sites


Abundance

0

100

200

300

400

500

B

Midhself sites


Abundance

0

100

200

300

400

500

C

Outer shelf sites


Abundance

500

1000

1500

2000

FIGURE 95. Total fish abundance across study sites: A) Inshore reefs; B) Mid-shelf reefs; and


10-15 m; IV= 15-20 m.


97

A

Inshore sites


H'n

0.0

0.5

1.0

1.5

2.0

2.5

3.0

B

Misdshelf sites


H'n

0.0

0.5

1.0

1.5

2.0

2.5

3.0

C

Outer shelf sites


H'n

0.0

0.5

1.0

1.5

2.0

2.5

3.0

FIGURE. 96. Fish species diversity (H’n) across study sites: A) Inshore reefs; B) Mid-shelf

reefs; and C) Outer shelf reefs. Mean±95% confidence interval. I= <5 m, II= 5-10

m, III= 10-15 m; IV= 15-20 m.


98

A

Inshore sites


J'n

0.0

0.2

0.4

0.6

0.8

1.0

B

Midshelf sites


J'n

0.0

0.2

0.4

0.6

0.8

1.0

C

Outer shelf sites


J'n

0.0

0.2

0.4

0.6

0.8

1.0

FIGURE. 97. Fish species evenness (J’n) across study sites: A) Inshore reefs; B) Mid-shelf


m, III= 10-15 m; IV= 15-20 m.


99

Mean total herbivore fish abundance averaged 76.7 individuals/count across inshore sites, with a

minimum of 13/count at the 10-15 m depth zone of EME and a maximum of 177/count at the <5

m depth zone of LAM (Figure 98). Total herbivore abundance across mid-shelf sites averaged

89 individuals/count, with a minimum of 50/count at the 10-15 m depth zone of RES and a

maximum of 131 individuals/count at the 5-10 m depth zone of CON. Total herbivores averaged

42 individuals/count across outer shelf sites, with a minimum of 16 individuals/count at the 10-

15 m depth zone of PPS and a maximum of 70 individuals/count at the 10-15 m depth zone of

GAL. Overall, total herbivore abundance declined with declining algal cover and distance from

the shore. Scrapers were the most abundant functional herbivore guild, followed by browsers and

non-denuders (Figure 99).

Mean total carnivore fish abundance showed a significant increase with increased distance from

the shore, averaging 13 individuals/count across inshore sites, with a minimum of 10/count at

the 5-10 m depth zone of RAT and a maximum of 19/count at the <5 m depth zone of LAM

(Figure 100). Total carnivore abundance across mid-shelf sites averaged 44 individuals/count,

with a minimum of 25 individuals/count at the <5 m depth zone of CON and a maximum of 110

individuals/count at the 10-15 m depth zone of RON. Total carnivores averaged 239

individuals/count across outer shelf sites, with a minimum of 43 individuals/count at the 10-15 m

depth zone of GAL and a maximum of 805/count at the 10-15 m depth zone of NEG.

Total carnivore guilds showed a significant difference in their assemblages depending on the

geographic location (Figure 101). For instance, generalist carnivores were proportionately

dominant across inshore and mid-shelf sites, while planktivores became more common across

mid-shelf sites, and eventually dominated outer shelf sites. Piscivore abundance was particularly

concerning. Piscivores averaged only 2.7 individuals/count across inshore sites, with a minimum

value of 0.8 individuals/count at the <5 m depth zone of EME and a maximum value of 5.5

individuals/count at the 5-10 m depth zone of EME. Piscivores averaged 3.1 individuals/count

across mid-shelf sites, with a minimum of 0.5 individuals/count at the <5 m depth zone of RES,

and a maximum of 16 individuals/count at the 10-15 m depth zone of RON. Piscivores averaged

10 individuals/count across outer shelf sites, with none present across the <5 m depth zone of

NEG, and a maximum of 64 individuals/count across the 15-20 m deep wall of PPS, mostly the


100

A

Inshore sites


He

rbiv

ore

ab

un

da

nce

0

100

200

300

400

B

Midshelf sites


He

rbiv

ore

ab

un

da

nce

0

100

200

300

400

C

Outer shelf sites


Herb

ivore

abundance

0

100

200

300

400

FIGURE 98. Total herbivore abundance across study sites: A) Inshore reefs; B) Mid-shelf


m, III= 10-15 m; IV= 15-20 m.


101

A

Inshore sites


He

rbiv

ore

ab

un

da

nce

0

20

40

60

80

100

120

140

160

180

200

Scrapers

Browsers

Non-denuders

B

Midshelf sites


He

rbiv

ore

ab

un

da

nce

0

20

40

60

80

100

120

140

160

180

200

C

Outer shelf sites


He

rbiv

ore

ab

un

da

nce

0

20

40

60

80

100

120

140

160

180

200

FIGURE 99. Mean herbivore guilds abundance across study sites: A) Inshore reefs; B) Mid-

shelf reefs; and C) Outer shelf reefs. Scrapers (Scaridae), Browsers

(Acanthuridae), Non-denuders (Pomacentridae). I= <5 m, II= 5-10 m, III= 10-15

m; IV= 15-20 m.


102

A

Inshore sites


To

tal ca

rniv

ore

ab

un

da

nce

0

20

40

60

80

100

120

140

160

B

Midshelf site


To

tal ca

rniv

ore

ab

un

da

nce

0

20

40

60

80

100

120

140

160

C

Outer shelf sites


To

tal ca

rniv

ore

ab

un

da

nce

0

200

400

600

800

1000

1200

1400

FIGURE 100. Total carnivore abundance across study sites: A) Inshore reefs; B) Mid-shelf


m, III= 10-15 m; IV= 15-20 m.


103

A

Inshore sites


Ca

rniv

ore

ab

un

da

nce

0

20

40

60

80

100

120

Generalists

Piscivores

Planktivores

B

Midshelf sites


Ca

rniv

ore

ab

un

da

nce

0

20

40

60

80

100

120

C

Outer shelf sites


Ca

rniv

ore

ab

un

da

nce

0

200

400

600

800

1000

FIGURE 101. Mean carnivore guilds abundance across study sites: A) Inshore reefs; B) Mid-

shelf reefs; and C) Outer shelf reefs. I= <5 m, II= 5-10 m, III= 10-15 m; IV= 15-

20 m.


104

Yellowtail snapper, Ocyurus chrysurus. These values are very low and share light on possibly

significant cross-shelf fishing impacts.

Mean omnivore fish abundance showed a significant increase with increased distance from the

shore, averaging less than 3 individuals/count across inshore sites, with none present at the 10-

15 m depth zone of EME and a maximum of 4/count at the <5 m depth zone of GUA (Figure

102). Omnivore abundance across mid-shelf sites averaged 33 individuals/count, with a

minimum of 4/count at the 10-15 m depth zone of RES, and the <5 m depth zone of CON and

RON, and a maximum of 121/count at the 10-15 m depth zone of RON. Omnivores averaged

127 individuals/count across outer shelf sites, with none present at the 10-15 m depth zone of

PPS and a maximum of 501/count at the 5-10 m depth zone of NEG.

There was a significant increase in mean total fish biomass with increasing distance from the

shore along the documented LBSP gradient (Figure 103). Total biomass averaged 3,222 g/count

across inshore sites, with a minimum of 1,329 g/count at the <5 m depth zone of EME and a

maximum of 6,622 g/count at the <5 m depth zone of LAM, mostly juvenile and semi-adult

parrotfishes (Scaridae). Total biomass increased to 5,221 g/count across mid-shelf sites, with a

minimum of 2,325 g/count at the 5-10 m depth zone of CON and a maximum of 19,291 g/count

at the 10-15 m depth zone of RON. Mean total biomass further increased to 15,110 g/count

across outer shelf sites, with a minimum of 2,099 g/count at the <5 m depth zone of NEG and a

maximum of 51,284 g/count at the 10-15 m depth zone of NEG. The lowest biomass values were

documented across the most degraded reef sites.

Mean total herbivore fish biomass averaged 2,392 g/count across inshore sites or a mean 74% of

the total observed biomass, with a minimum of 503 g/count at the 10-15 m depth zone of EME

and a maximum of 5,550 g/count at the <5 m depth zone of LAM (Figure 104). Total herbivore

biomass across mid-shelf sites averaged 2,357 g/count or a mean 45% of the total observed

biomass, with a minimum of 1,363 g/count at the 10-15 m depth zone of CON and a maximum

of 3,060 g/count at the 5-10 m depth zone of RON. Total herbivore biomass averaged 2,941

g/count across outer shelf sites or 19% of the total documented biomass, with a minimum of 689

g/count at the 10-15 m depth zone of PPS and a maximum of 8,104 g/count at the 5-10 m depth

zone of NEG. Overall, total herbivore biomass declined with declining algal cover and distance

from the shore. Herbivore proportions also declined accordingly. Scraper herbivores were the


105

A

Inshore sites


Om

niv

ore

ab

un

da

nce

0

50

100

150

200

250

B

Midshelf sites


Om

niv

ore

ab

un

da

nce

0

50

100

150

200

250

C

Outer shelf sites


Lo

g1

0 O

mn

ivo

re a

bu

nd

an

ce

0.1

1

10

100

1000

FIGURE 102. Omnivore abundance across study sites: A) Inshore reefs; B) Mid-shelf reefs; and


10-15 m; IV= 15-20 m.


106

A

Inshore sites


To

tal b

iom

ass (

g)

0

10000

20000

30000

40000

50000

60000

B

Midshelf sites


To

tal b

iom

ass (

g)

0

10000

20000

30000

40000

50000

60000

C

Outer shelf sites


To

tal b

iom

ass (

g)

0

10000

20000

30000

40000

50000

60000

FIGURE 103. Total fish biomass across study sites: A) Inshore reefs; B) Mid-shelf reefs; and C)

Outer shelf reefs. Mean±95% confidence interval. I= <5 m, II= 5-10 m, III= 10-15

m; IV= 15-20 m.


107

A

Inshore sites


To

tal h

erb

ivo

re b

iom

ass (

g)

0

5000

10000

15000

20000

25000

B

Midshelf sites


To

tal h

erb

ivo

re b

iom

ass (

g)

0

5000

10000

15000

20000

25000

C

Outer shelf sites


To

tal h

erb

ivo

re b

iom

ass (

g)

0

5000

10000

15000

20000

25000

FIGURE 104. Total herbivore biomass across study sites: A) Inshore reefs; B) Mid-shelf reefs;

and C) Outer shelf reefs. Mean±95% confidence interval. I= <5 m, II= 5-10 m,

III= 10-15 m; IV= 15-20 m.


108

A

Inshore sites


Bio

ma

ss (

g)

0

2000

4000

6000

8000

10000

Scrapers

Browsers

Non-denuders

B

Midshelf sites


Bio

ma

ss (

g)

0

2000

4000

6000

8000

10000

C

Outer shelf sites


Bio

mass (

g)

0

2000

4000

6000

8000

10000

FIGURE 105. Mean herbivore guilds biomass across study sites: A) Inshore reefs; B) Mid-shelf

reefs; and C) Outer shelf reefs. Scrapers (Scaridae), Browsers (Acanthuridae),

Non-denuders (Pomacentridae). I= <5 m, II= 5-10 m, III= 10-15 m; IV= 15-20 m.


109

most abundant functional herbivore guild, followed by browsers and non-denuders, although

brower biomass increased with distance from the shore (Figure 105).

Mean total carnivore fish biomass showed a significant increase with increased distance from the

shore, averaging 809 g/count across inshore sites or 25% or the total observed biomass, with a

minimum of 83 g/count at the 5-10 m depth zone of OST and a maximum of 2,071 g/count at the

5-10 m depth zone of EME (Figure 106). Total carnivore biomass across mid-shelf sites

averaged 2,832 g/count or 54% or the total observed biomass, with a minimum of 210 g/count at

the <5 m depth zone of CON and a maximum of 16,660 g/count at the 10-15 m depth zone of

RON. Total carnivores averaged 11,870 g/count across outer shelf sites or 79% or the total

observed biomass, with a minimum of 600 g/count at the 5-10 m depth zone of NEG and a

maximum of 49,588 g/count at the 10-15 m depth zone of NEG.

Total carnivore guilds also showed a significant difference in their assemblages depending on the

geographic location (Figure 107). For instance, generalist carnivore biomass was proportionately

dominant across inshore and mid-shelf sites, while planktivores became more common across

mid-shelf sites, and eventually dominated outer shelf sites. Piscivore biomass was also

concerning. Piscivore biomass averaged only 446 g/count across inshore sites, with a minimum

value of less than 5g/count at the <5 m depth zone of OST and a maximum value of 1,853

g/count at the 5-10 m depth zone of EME. Piscivores averaged 1,400 g/count across mid-shelf

sites, with a minimum of 49 g/count at the <5 m depth zone of RON, and a maximum of 11,453

g/count at the 10-15 m depth zone of RON. Piscivores averaged 1,778 g/count across outer shelf

sites, with none present across the <5 m depth zone of NEG, and a maximum of 11,394 g/count

across the 15-20 m deep wall of PPS, mostly Ocyurus chrysurus. These values, however, are

very low and share light on possibly significant cross-shelf fishing impacts.

Omnivore biomass showed a significant increase with increased distance from the shore,

averaging 21 g/count across inshore sites, with none present at the 10-15 m depth zone of EME

and a maximum of 53 g/count at the <5 m depth zone of GUA (Figure 108). Omnivore biomass

across mid-shelf sites averaged 31 g/count, with a minimum of less than 10 g/count at the 10-15

m depth zone of CON and a maximum of 44 g/count at the <5 m depth zone of RES. Omnivores

averaged 34 g/count across outer shelf sites, with none present at the 10-15 m depth zone of PPS

and <5 m depth zone of GAL, and a maximum of 133 g/count at the 5-10 m depth zone of NEG.


110

A

Inshore sites


To

tal ca

rniv

ore

bio

ma

ss (

g)

0

5000

10000

15000

20000

25000

B

Midshelf sites


To

tal ca

rniv

ore

bio

ma

ss (

g)

0

5000

10000

15000

20000

25000

C

Outer shelf sites


To

tal ca

rniv

ore

bio

ma

ss (

g)

0

10000

20000

30000

40000

50000

60000

FIGURE 106. Total carnivore biomass across study sites: A) Inshore reefs; B) Mid-shelf reefs;

and C) Outer shelf reefs. Mean±95% confidence interval. I= <5 m, II= 5-10 m,

III= 10-15 m; IV= 15-20 m.


111

A

Inshore sites


Bio

ma

ss (

g)

0

10000

20000

30000

40000

50000

Generalists

Piscivores

Planktivores

B

Midshelf sites


Bio

ma

ss (

g)

0

10000

20000

30000

40000

50000

C

Outer shelf sites


Bio

ma

ss (

g)

0

10000

20000

30000

40000

50000

FIGURE 107. Mean carnivore guilds biomass across study sites: A) Inshore reefs; B) Mid-shelf

reefs; and C) Outer shelf reefs. I= <5 m, II= 5-10 m, III= 10-15 m; IV= 15-20 m.


112

A

Inshore sites


Om

niv

ore

bio

ma

ss (

g)

0

50

100

150

200

250

B

Midshelf sites


Om

niv

ore

bio

ma

ss (

g)

0

50

100

150

200

250

C

Outer shlef sites


Om

niv

ore

bio

ma

ss (

g)

0

50

100

150

200

250

FIGURE 108. Omnivore biomass across study sites: A) Inshore reefs; B) Mid-shelf reefs; and C)

Outer shelf reefs. Mean±95% confidence interval. I= <5 m, II= 5-10 m, III= 10-15

m; IV= 15-20 m.


113

An assessment of selected coral reef fish taxa showed evidence of a potential combination of

long-term cross-shelf fishing impacts, depth and reef structural and location effects. Edible

grouper subfamily Epinephelinae showed very low abundances across the entire shelf, with a

mean Red Hind (Epinephelus guttatus) abundance of 0.07 individuals/count across inshore reefs,

0.33/count across mid-shelf sites, and 0.19/count across outer shelf sites (Figure 109). No Rock

hind (E. adscensionis) were observed across inshore sites, 0.06/count was documented across

mid-shelf sites, and 0.19/count across outer shelf sites. Less than 0.04 individuals/count of

Graysby (Cephalopholis cruentata) were observed across inshore sites, less than 0.03/count

across mid-shelf sites, and 0.13/count across outer shelf sites. No individuals of Coney (C. fulva)

were observed across in shore sites, 0.22/count was observed across mid-shelf sites, and

0.44/count across outer shelf sites. Grouper abundance showed a general trend of increase with

increasing distance, but mean abundances were overall very low (less than 1 individual/count),

which implies potential chronic fishing impacts. The other interesting observations were that

Epinehelinae assemblages showed significant variability across sites, and that E. adscencionis

and C. fulva were completely absent across inshore sites.

Snappers (Lutjanidae) species composition also showed remarkable variability across sites, with

increasing abundance with increasing distance from the shore (Figure 110). Mean Gray snapper

(Lutjanus griseus) abundance was 0.04 individuals/count across inshore reefs, 0.03/count across

mid-shelf sites, and none across outer shelf sites. Schoolmaster (L. apodus) abundance averaged

0.22/count across inshore sites, 0.31/count was documented across mid-shelf sites, and

0.59/count across outer shelf sites. Only 0.05 individuals/count of Dog snapper (L. jocu) were

observed across inshore sites, none across mid-shelf sites, and 0.03/count across outer shelf sites.

Mutton snapper (L. analis) showed an abundance of 0.04/count across in shore sites, none across

mid-shelf sites, and 0.03/count across outer shelf sites. Lane snapper (L. synagris) showed an

abundance of only 0.01/count across in shore sites, 0.03/count was observed across mid-shelf

sites, and none across outer shelf sites. Mahogany snapper (L. mahogoni) showed an abundance

of 0.06/count across in shore sites, none across mid-shelf sites, and none across outer shelf sites.

Finally, Yellowtail snapper (Ocyurus chrysurus) showed an abundance of 1.25/count across in

shore sites, with 0.89/count across mid-shelf sites, and 4.09/count across outer shelf sites. The

abundance of O. chrysurus and L. apodus showed a general trend of increase with increasing

distance.


114

A

Inshore sites


Ep

ine

ph

elin

ae

0.0

0.5

1.0

1.5

2.0

2.5

E. guttatus

E. adsencionis

C. cruentata

C. fulva

B

Midshelf sites


Ep

ine

ph

elin

ae

0.0

0.5

1.0

1.5

2.0

2.5

C

Outer shelf sites


Ep

ine

ph

elin

ae

0.0

0.5

1.0

1.5

2.0

2.5

FIGURE 109. Edible groupers (sub-family Epinephelinae) abundance across study sites: A)

Inshore reefs; B) Mid-shelf reefs; and C) Outer shelf reefs. Mean±95%

confidence interval. I= <5 m, II= 5-10 m, III= 10-15 m; IV= 15-20 m.


115

Snappers (Lutjanidae) species composition also showed remarkable variability across sites, with

increasing abundance with increasing distance from the shore (Figure 110). Mean Gray snapper

(Lutjanus griseus) abundance was 0.04 individuals/count across inshore reefs, 0.03/count across

mid-shelf sites, and none across outer shelf sites. Schoolmaster (L. apodus) abundance averaged

0.22/count across inshore sites, 0.31/count was documented across mid-shelf sites, and

0.59/count across outer shelf sites. Only 0.05 individuals/count of Dog snapper (L. jocu) were

observed across inshore sites, none across mid-shelf sites, and 0.03/count across outer shelf sites.

Mutton snapper (L. analis) showed an abundance of 0.04/count across in shore sites, none across

mid-shelf sites, and 0.03/count across outer shelf sites. Lane snapper (L. synagris) showed an

abundance of only 0.01/count across inshore sites, 0.03/count was observed across mid-shelf

sites, and none across outer shelf sites. Mahogany snapper (L. mahogoni) showed an abundance

of 0.06/count across in shore sites, none across mid-shelf sites, and none across outer shelf sites.

Finally, Yellowtail snapper (Ocyurus chrysurus) showed an abundance of 1.25/count across

inshore sites, with 0.89/count across mid-shelf sites, and 4.09/count across outer shelf sites. The

abundance of O. chrysurus and L. apodus showed a general trend of increase with increasing

distance.

Grunts (Haemulidae) species composition showed remarkable variability across sites, but mostly

depending on depth and reef physical structure (Figure 111). Tomtate (Haemulon aurolineatum)

abundance was 0.17 individuals/count across inshore reefs, 0.67/count across mid-shelf sites, and

5.0/count across outer shelf sites. Smallmouth grunt (H. chrysargyreum) abundance averaged

0.29/count across inshore sites, and none across mid-shelf and outer shelf sites. No individuals of

Sailors choice (H. parrai) were observed across inshore sites, 0.03/count across mid-shelf sites,

and none across outer shelf sites. Spanish grunt (H. macrostomum) showed an abundance of

0.02/count across in shore sites, 0.11/count across mid-shelf sites, and none across outer shelf

sites. Margate (H. album) was absent across in shore sites, but had 0.03/count across mid-shelf

sites, and none across outer shelf sites. White grunt (H. plumieri) showed an abundance of

0.24/count across inshore sites, 0.47/count across mid-shelf sites, and 0.06/count across outer

shelf sites. Caesar grunt (H. carbonarium) was also absent across inshore sites, but had

0.14/count across mid-shelf sites, and 0.19/count across outer shelf sites. Bluestriped grunt (H.

sciurus) showed an abundance of 0.02/count across inshore sites, and none across mid-shelf and

outer shelf sites. Finally, French grunt (H. flavolineatum) showed an abundance of 0.38/count


116

A

Inshore sites


Lu

tja

nid

ae

0

1

2

3

4

5

L. griseus

L. apodus

L. jocu

L. analis

L. synagris

L. mahogoni

O. chrysurus

B

Midshelf sites


Lu

tja

nid

ae

0

1

2

3

4

5

C

Outer shelf sites


Lu

tja

nid

ae

0

5

10

15

20

25

FIGURE 110. Snappers (family Lutjanidae) abundance across study sites: A) Inshore reefs; B)

Mid-shelf reefs; and C) Outer shelf reefs. Mean±95% confidence interval. I= <5

m, II= 5-10 m, III= 10-15 m; IV= 15-20 m.


117

A

Inshore sites


Ha

em

ulid

ae

0

1

2

3

4

5

6

H. aurolineatum

H. chrysargyreum

H. parrai

H. macrostomum

H. album

H. plumieri

H. carbonarium

H. sciurus

H. flavolineatum

B

Midshelf sites


Ha

em

ulid

ae

0

1

2

3

4

5

6

C

Outer shelf sites


Ha

em

ulid

ae

0

10

20

30

40

FIGURE 111. Grunts (family Haemulidae) abundance across study sites: A) Inshore reefs; B)


m, II= 5-10 m, III= 10-15 m; IV= 15-20 m.


118

across inshore sites, with 0.19/count across mid-shelf sites, and 0.06/count across outer shelf

sites. The abundance of H. aurolineatum and H. carbonarium showed a general trend of increase

with increasing distance. All other species showed a spatial variation trend which seemed to be

more related to a combination of depth and reef benthic spatial heterogeneity across shallower

zones.

Parrotfishes (Scaridae) were the most abundant taxa across sites, but their species composition

showed variable assemblages across sites (Figure 112). Yellowtail parrotfish (Sparisoma

rubripinne) abundance was 15.18 individuals/count across inshore reefs, 1.11/count across mid-

shelf sites, and 1.47/count across outer shelf sites. Redtail parrotfish (Sp. chrysopterum)

abundance averaged 0.08/count across inshore sites, 0.47/count was documented across mid-

shelf sites, and 0.22/count across outer shelf sites. Stoplight parrotfish (Sp. viride) averaged

4.16/count across inshore sites, 11.06/count across mid-shelf sites, and 3.16/count across outer

shelf sites. Redband parrotfish (S. aurofrenatum) showed an abundance of 2.35/count across

inshore sites, 12.89/count across mid-shelf sites, and 3.78/count across outer shelf sites.

Bucktooth parrotfish (Sp. radians) showed an abundance of only 0.28/count across in shore sites,

0.03/count was observed across mid-shelf sites, and was absent across outer shelf sites.

Greenblotch parrotfish (Sp. atomarium) was absent across inshore sites, had 1.92/count across

mid-shelf sites, and was also absent across outer shelf sites. Queen parrotfish (Scarus vetula) had

an abundance of 0.19/count across inshore sites, 0.56/count across mid-shelf sites, and

0.78/count across outer shelf sites. Striped parrotfish (Sc. iserti) averaged 23.07/count across

inshore sites, 34.44/count across mid-shelf sites, and 12.69/count across outer shelf sites.

Princess parrotfish (Sc. taeniopterus) averaged 14.42/count across inshore sites, 5.53/count

across mid-shelf sites, and 2.88/count across outer shelf sites. Rainbow parrotfish (Sc.

guacamaia) was absent across inshore and mid-shelf sites, but averaged 0.19/count across outer

shelf sites. Finally, Bluelip (Cryptotomus roseus) showed an abundance of 0.14/count across

inshore sites, and was absent across mid-shelf and outer shelf sites. Parrotfish distribution was

largely influenced by depth and distance from the shore, in particular for species such as Sp.

rubripinne, Sc. iserti, and Sc. taeniopterus. Shallow, degraded inshore reefs such (OST, LAM,

RAT still have a critical function as nursery grounds and support very high densities of juvenile

and semi-adult individuals. Also, Scarid abundance seems to be affected by algal distribution

patterns. Species vulnerable to fishing (Sc. vetula, Sc. guacamaia) showed a limited distribution.


119

A

Inshore sites


Sca

rid

ae

0

20

40

60

80

100

120

140

160

Sp. rubripinne

Sp. chrysopterum

Sp. viride

Sp. aurofrenatum

Sp. radians

Sp. atomarium

Sc. vetula

Sc. iserti

Sc. taeniopterus

Sc. guacamaia

C. roseus

B

Midshelf sites


Sca

rid

ae

0

20

40

60

80

100

120

140

160

C

Outer shelf sites


Sca

rid

ae

0

20

40

60

80

100

120

140

160

FIGURE 112. Parrotfishes (family Scaridae) abundance across study sites: A) Inshore reefs; B)


m, II= 5-10 m, III= 10-15 m; IV= 15-20 m.


120

The abundance of butterflyfishes (Chaetodontidae) was overall low and showed slight variation

depending on overall reef condition (Figure 113). Foureye butterflyfish (Chaetodon capistratus)

abundance was 1.10 individuals/count across inshore reefs, 1.47/count across mid-shelf sites, and

0.88/count across outer shelf sites. Banded butterflyfish (C. striatus) abundance averaged

0.18/count across inshore sites, 0.08/count was documented across mid-shelf sites, and was

absent across outer shelf sites. Spotfin butterflyfish (C. ocellatus) averaged 0.07/count across

inshore sites, had none across mid-shelf sites, and averaged 0.06/count across outer shelf sites.

Finally, Reef butterflyfish (C. sedentarius) was absent across inshore sites, and had 0.03/count

across mid-shelf sites, and 0.19/count across outer shelf sites.

The abundance of damselfish (Pomacentridae) showed significant species-specific spatial

variability (Figure 114). The Brazilian damselfish (Stegastes fuscus) abundance was 0.10

individuals/count across inshore reefs, and was absent across mid-shelf and outer shelf sites.

Threespot damselfish (S. planifrons) abundance averaged 4.69/count across inshore sites,

6.78/count across mid-shelf sites, and 5.47/count across outer shelf sites. Cocoa damselfish (S.

variabilis) averaged 0.07/count across inshore sites, 1.53/count across mid-shelf sites, and

0.34/count across outer shelf sites. Beaugregory (S. leucostictus) averaged 1.36/count across

inshore sites, 0.03/count across mid-shelf sites, and 0.40/count across outer shelf sites. Bicolor

damselfish (S. partitus) averaged 0.04/count across inshore sites, 2.39/count across mid-shelf

sites, and 8.53/count across outer shelf sites. Dusky damselfish (S. adustus) averaged 0.82/count

across inshore sites, 0.67/count across mid-shelf sites, and 0.06/count across outer shelf sites.

Longfin damselfish (S. diencaeus) averaged 0.82/count across inshore sites, 0.61/count across

mid-shelf sites, and 0.94/count across outer shelf sites. Finally, Yellowtail damselfish

(Microspathodon chrysurus) averaged 0.20/count across inshore sites, 1.11/count across mid-

shelf sites, and 2.66/count across outer shelf sites. Stegastes planifrons showed consistently high

densities across the shelf, regardless of depth and distance from the shore. But species such as S.

leucostictus and S. adustus were more abundant across shallow inshore sites, while species such

as S. partitus and M. chrysurus were more abundant across outer shelf and deeper sites exposed

to stronger circulation.


121

A

Inshore sites


Ch

ae

tod

on

tid

ae

0

1

2

3

4

5

6

C. capistratus

C. striatus

C. ocellatus

C. sedentarius

B

Midshelf sites


Ch

ae

tod

on

tid

ae

0

1

2

3

4

5

6

C

Outer shelf sites


Ch

ae

tod

on

tid

ae

0

1

2

3

4

5

6

FIGURE 113. Butterflyfishes (family Chaetodontidae) abundance across study sites: A) Inshore

reefs; B) Mid-shelf reefs; and C) Outer shelf reefs. Mean±95% confidence

interval. I= <5 m, II= 5-10 m, III= 10-15 m; IV= 15-20 m.


122

A

Inshore sites


Po

ma

ce

ntr

ida

e

0

5

10

15

20

25

30

35

S. fuscus

S. planifrons

S. variabilis

S. leucostictus

S. partitus

S. adustus

S. diencaeus

M. chrysurus

B

Midshelf sites


Po

ma

ce

ntr

ida

e

0

5

10

15

20

25

30

35

C

Outer shelf sites


Po

ma

ce

ntr

ida

e

0

5

10

15

20

25

30

35

FIGURE 114. Damselfishes (family Pomacentridae) abundance across study sites: A) Inshore

reefs; B) Mid-shelf reefs; and C) Outer shelf reefs. Mean±95% confidence

interval. I= <5 m, II= 5-10 m, III= 10-15 m; IV= 15-20 m.


123

IV. DISCUSION

a. LBSP cross-shelf spatial gradient

This study showed important evidence that there were signs of a LBSP gradient across the

western Puerto Rico shelf and that chronic water quality decline has significantly affected the

face of coral reef benthic communities, coral recruit assemblages, and have indirectly affected

reef fish communities. A snapshot view of LBSP showed that particularly turbidity and PO4 and

NH4+ concentrations increased along inshore sites, while dissolved oxygen concentration

declined along inshore waters. It is particularly concerning that EME reef site and to some extent

GUA are being exposed to recurrent pulses of sewage effluents from four different sewage

treatment facilities at Boquerón Bay. Elevated PO4, NH4+, chlorophyll-a, and OABs

concentrations suggest at EME suggest that tidal cycles may continuously expose coral reef

adjacent to Boquerón Bay to recurrent sewage pollution and eutrophication impacts. In the other

hand, turbidity, OABs resulted fairly high at other inshore sites such as JOY, RAT, and OST.

Their proximity to Joyuda Bay and Puerto Real Bay continuously expose these sampling sites to

recurrent polluted water pulses. But a particular concern was water quality conditions across

outer shelf sites, where dissolved nutrient concentrations and chlorophyll-a concentrations

exceeded recommended levels for healthy coral reefs. Pollution across outer shelf sites may also

come from other significant sources such as the Río Guanajibo, Río Yagüez, and the Mayagüez

Bay.

Previous studies have suggested that PO4 concentration on coral reef habitats should not exceed

0.3 μM (Lapointe and Clark, 1992), and that any concentration above 1.5 μM would be too high

for coral reefs (Lapointe and Matzie, 1996). Our study showed mean PO4 concentration of 3.46

μM across inshore sites, of 2.67 μM across mid-shelf sites, and of 1.23 μM across outer shelf

sites. These values resulted 10.5 times above the recommended concentration for healthy coral

reefs across inshore sites, 7.9 times across mid-shelf sites, and 3.1 times across outer shelf sites.

These results suggest that PO4 pollution pulses are a concerning shelf-wide scale phenomenon.

Previous studies in Puerto Rico have also shown previous high PO4 concentrations significantly

impacting coral assemblages along the northern coast (Díaz-Ortega and Hernández-Delgado,

2014).


124

NH4+ concentrations above 24 μM have been deemed as too high, when the recommended

concentration for coral reefs should not exceed 0.1 μM (Lapointe and Clark, 1992). We obtained

mean NH4+ concentrations that ranged of 98.09 μM across inshore sites, with even an alarming

260 μM mean value at EME during an ebbing tide. Mean NH4+ concentration was 59.09 μM

across mid-shelf sites, and 15.63 μM across outer shelf sites. Observed NH4+ concentrations

across inshore sites exceeded 981 times the recommended concentration for healthy coral reefs.

The observed NH4+ mean concentration at EME, which is located just at the mouth of Boquerón

Bay, exceeded by 2,599 times the recommended limits for healthy coral reefs. Concentrations

also exceeded the recommended limits 590 times across mid-shelf sites, and 155 times across

outer shelf sites. Such levels were also attained in previous studies across the Vega Baja coast in

northern Puerto Rico, where coral reefs are rapidly declining (Díaz-Ortega and Hernández-

Delgado, 2014). But the spatial scale of the observed patterns in this study seems to be

unprecedented.

Another key concern associated to sewage and eutrophication impacts was the elevated cross-

shelf mean chlorophyll-a concentration range found in this study in comparison to the

recommended concentration for coral reef waters of 0.3-0.5 μg/L (Lapointe and Clark, 1992;

Otero, 2009). Mean chlorophyll-a concentration was 2.38 μg/L across inshore sites, 2.07 μg/L

across mid-shelf sites, and 2.12 μg/L across outer shelf sites. These values were up 3.8 to 6.9

times higher than recommended limits for healthy coral reefs across inshore sites, 3.1 to 5.9

times across mid-shelf sites, and 3.2 to 6.1 times across outer shelf sites.

Mean dissolved oxygen concentration in this study was also highly concerning, when any

concentration below 2.5 mg/L is considered hypoxic (Lapointe and Matzie, 1996). Mean

dissolved oxygen concentration was 4.12 mg/L across inshore sites, 4.09 mg/L across mid-shelf

sites, and 5.63 mg/L across outer shelf sites. These values, although far from hypoxic levels,

were still concerning as we consider them low for open waters across the shelf and may evidence

the chronic state and the large spatial scale of water quality decline as a result of LBSP.

Water quality data in this study was fairly limited as we had no temporal replication.

Nonetheless, PCO analysis of observed spatial patterns of water quality explained 74.3% of the

variation, confirming observed clustering pattern among inshore, mid-shelf and outer shelf sites


125

across a LBSP gradient. Documented spatial patterns were also highly consistent with findings of

cross-shelf scale pollution patterns documented by Bonkosky et al. (2009). Turbidity patterns

were also consistent with previous unpublished observations from year 2000 (Hernández-

Delgado, unpub. data). Therefore, it was reasonable to assume that observed spatial patterns of

water quality conditions in this study were highly consistent with chronic large-scale degradation

at least over the last two decades and that the observed LBSP stress gradient in the form of

chronic turbidity and eutrophication, mostly associated to sewage pollution, represent a nearly

permanent state.

b. Coral reef benthic community structure

Benthic community structure showed a highly significant variation along the observed LBSP

stress gradient. Coral species richness, H’n, J’n and percent living cover increased with

increasing distance from LBSP. In contrast, overall algal cover increased across inshore and

some of the mid-shelf sites. Further, algal taxonomic composition also showed a significant

variation across the documented water quality stress gradient. PCO analysis explained 56.4% of

the spatial variation in benthic community structure and showed a spatial clustering pattern of

surveyed sites that closely resembled spatial clustering patterns associated to the documented

LBSP gradient. Turbidity and the concentration of PO4 and NH4+ were the most significant water

quality parameters that explained spatial patterns in benthic community structure. LINKTREE

analysis also showed that the combination of high NH4+ concentration and high turbidity was the

most influencing factor on the observed spatial variation in benthic community structure. The

rest of the observed variation in benthic community structure could have resulted from natural

variability associated to coral reef spatial heterogeneity, depth and local oceanographic

circulation patterns.

c. Coral recruit community structure

Similarly, coral recruit community structure was significantly influenced by the observed LBSP

spatial gradient. Coral recruit abundance increased with increasing distance from LBSP. Also,

species assemblages showed significant variation with increasing distance from the shore.

Further, recruit abundance increased with increasing dissolved oxygen concentration, and

declined with increasing turbidity and NH4+ concentration. PCO analysis explained 61% of the


126

spatial variation observed in coral recruit assemblages, mostly as a result of influences by

turbidity, NH4+ and PO4. LINKTREE analysis confirmed that the combination of high turbidity

and low dissolved oxygen concentration was the most influencing condition affecting the spatial

distribution of coral recruits. Other factors could have included coral reef spatial heterogeneity,

depth and local oceanographic circulation patterns. Nevertheless, observed patterns were pretty

revealing and concerning as the lack of abundant coral recruitment, particularly of large reef-

building species implies, with few exceptions across outer shelf sites, a widespread natural lack

of recovery potential from disturbance across inshore and most of the surveyed mid-shelf coral

reef sites.

d. Coral reef fish community structure

Coral reef fish community structure also showed a highly significant cross-shelf spatial variation

influenced by a combination of factors. There was a significant general increase in species

richness, total abundance and biomass, H’n, J’n, total carnivore abundance and biomass,

piscivore abundance and biomass, and omnivore abundance and biomass with increasing

distance from the shore. PCO analysis explained 48% of the observed spatial patterns, mostly as

a result of the spatial variation in the concentration of NH4+ and PO4, and in turbidity.

LINKTREE analysis suggested that the combination between high PO4 and OABs

concentrations explained most of the observed variation in fish community structure across the

observed LBSP gradient. These findings were highly consistent with previous observations of

Bejarano and Appeldoorn (2013) that found declining fish community assemblages across turbid

environments. The rest of the observed variation in this study could have resulted from

influences by local factors such as depth, benthic spatial heterogeneity, reef condition and

position, and local oceanographic circulation patterns.

Four additional important findings were documented in this study. First, the distribution of many

selected fishery-targeted species showed very low cross-shelf abundances, though many of them

showed an increasing abundance with increasing distance from the shore. This may imply the

combination of potential LBSP impacts but moreover chronic fishing impacts, particularly across

inshore and mid-shelf sites. That was the case of members of the edible groupers in subfamily

Epinephelinae and snappers (Lutjanidae). The second important finding is that the abundance of


127

potential indicator species, such as butterflyfishes (Chaetodontidae), the Yellowtail damselfish

(Microspathodon chrysurus), and the Bicolor damselfish (Stegastes partitus) increased with

increased distance from the shore, across the documented LBSP spatial patterns. Other cryptic

fish species also increased along the LBSP gradient, with many completely absent across inshore

sites (data not shown). These findings were also consistent with observations from Bejarano and

Appeldoorn (2013) and suggest potential indirect LBSP effects on fish community structure

mediated by previos chronic benthic habitat degradation and habitat environmental quality

decline.

A third important finding was the dominant state of parrotfishes (Scaridae) across most of the

surveyed sites. Parrotfishes were the most important taxa explained variability across sites, but

juvenile and semi-adult stages were particularly dominant and abundant across shallow, highly-

degraded inshore reef sites. This finding has significant implications for the maintenance of

regional fisheries as it suggest that, regardless of their coral-depleted state, inshore, turbid and

polluted reef habitats still have a critical role as nursery ground for a keystone functional group

of scraper herbivores. The fact that we also documented the presence of juvenile and semi-adult

stages of several species of snappers (Lutjanidae) and grunts (Haemulidae) across inshore sites

stressed out the importance of shallow inshore reef sites as a key component of the ontogenic

changes of many fish species. This implies the critical need to: 1) provide the most strong

protection to shallow inshore coral reef communities and adjacent habitats; 2) strengthen

measures to immediately implement the best management practices (BMPs) possible to reduce

LBSP to adjacent coastal waters; 3) to implement long-term ecological monitoring efforts to

address ecological spatio-temporal changes in benthic community structure and coral recruitment

rates across inshore reef sites; 4) to also address spatio-temporal changes in fish community

structure across inshore sites, as well as across other cross-shelf sites in order to improve our

understanding of local spatio-temporal connectivity among sites and ontogenic shifts in many

commercially-important fish species. Therefore, existing large-scale LBSP stress represents a

major concern for the maintenance and natural recovery ability of coral reef fish communities.

Critical habitats necessary for natural ontogenic shifts of multiple species are severely degraded

and water quality still remains significantly compromised. Also, large-scale influences of

pollution from large rivers and from Mayagüez Bay also affect mid-shelf and outer shelf sites

across Arrecifes Tourmaline Natural Reserve. Potential effects of LBSP on the long-term


128

benefits of seasonal fishing closures within the natural reserve still remain largely unknown and

must also be addressed.

e. Implications for coral reef functions

Observed LBSP across the western Puerto Rico shelf in the form of high turbidity,

eutrophication and sewage pollution is a concern for coral reef resilience and for the

conservation and recovery of coral reef functions.

Turbidity, sewage pollution and eutrophication across coastal waters were not only confined to

the immediate vicinity of receiving waters and have been shown to potentially extend with tidal

and wind-driven currents through large geographic areas across the entire shelf (Bonkosky et al.,

2009), negatively impacting extensive coral reef communities (Hernández-Delgado et al., 2010).

Similar observations were made across the Vega Baja coast (Hernández-Delgado et al., 2011a,b;

Díaz-Ortega and Hernández_Delgado, 2014). Sewage impacts are often associated to

eutrophication and turbidity (Pastorok and Bilyard, 1985; Lapointe and Clark, 1992; Lapointe

and Matzie, 1996; Cloern, 2001; Díaz-Ortega and Hernández_Delgado, 2014). It has also been

associated to altered coral-associated microbial community composition (Sekar et al., 2008), to

an increased prevalence coral diseases (Patterson et al., 2002; Kaczmarsky et al., 2005), and to

declining coral survival rates (Walker and Ormond, 1982; Pastorok and Bilyard, 1985).

Declining coral skeletal extension rates (Tomascik and Sander, 1985), reproductive potential

(Tomascik and Sander, 1987), and larval settlement rates (Tomascik, 1991) can also occur under

eutrophied/polluted conditions. If chronic, these factors can negatively affect the natural coral

reef recovery ability from disturbance. In the long run, it can affect ecosystem resilience,

functions, services, and benefits.

Sewage and eutrophication impacts often result in a combination of system- and species-specific

responses, as well as cascading direct and indirect effects that could result in major long-term

phase shifts in benthic community structure, favoring dominance by fleshy macroalgae and non-

reef building taxa (Hernández-Delgado et al., 2010, 2011a). Such phase shifts could be

irreversible in long-term scales (Hughes, 1994). Sewage-associated eutrophication impacts can

also result in an accelerated reef decline often due to a combination of synergistic impacts,

mostly from sediments and turbidity (Meesters et al., 1998; Cloern, 2001; Szmant, 2002). This


129

combination of factors and their long-term shift from pulse to chronic events may explain why

some coral reefs across the entire western Puerto Rico shelf have shown the observed declines.

V. CONCLUSIONS

Coral reef ecosystems across the entire western Puerto Rico shelf were showing unequivocal

signs of decline mostly as a result of chronic LBSP impacts. The state of decline is reflected both

in benthic community structure and in coral recruit assemblages. But also, impacts have

indirectly affected coral reef fish community structure. Evidence from this study supports the

long-term chronic nature of the observed state of environmental decline. Observed conditions are

biologically unsustainable for all of the surveyed inshore coral reefs and for most of the surveyed

mid-shelf reefs. Although surveyed outer shelf reefs were in better shape than inshore and that

most of the mid-shelf sites, they were also showing important signs of decline. Impacts by

declining coastal water quality and non-point source pollution across the western Puerto Rico

shelf has been extensively documented in the past (Bonkosky et al., 2009; Hernández-Delgado et

al., 2010), but has still received little attention from Commonwealth and U.S. Federal regulatory

agencies to prevent or minimize further damage to these unique resources. Observed

chlorophyll-a, PO4 and NH4+ concentrations are very far from sustainable and point out at the

paramount importance of implementing specific water quality legal limits for coral reef waters

for chlorophyll-a (0.3 μg/L), PO4 (0.3 μM), NH4+ (0.1 μM), and a ratio of PO4:Total N of 7:1 or

lower.

There is also an imperative need to eliminate illegal raw sewage dumping from septic tank

effluents, malfunctioning sewers, and other non-point sources from properties adjacent to coastal

waters a Boquerón Bay, Puerto Real Bay, Joyuda Bay, Bramadero Bay, and Mayagüez Bay, and

to Río Guanajibo and Río Yagüez watersheds. It would also be important to implement BMPs to

foster an improved management of soil erosion, runoff and sedimentation of small-scale

watersheds across coastal adjacent lands. This should also include improving sewer systems and

upgrading existing sewage treatment plant facilities, particularly at Boquerón Bay. There is also

a need to improve the implementation of erosion-sedimentation controls on any construction

activity across adjacent watersheds or coastal waters.


130

There is also a need to implement a permanent long-term ecological monitoring program for

coral reef benthic communities, coral recruitment rates and reef fish communities, in

combination with a permanent, standard water quality monitoring program. We also encourage

testing methods aimed at coral propagation using local coral genetic clones which are naturally

resistant to local water quality conditions. Low-tech coral farming and reef rehabilitation have

become critical for helping coral reefs to recover from chronic disturbance.

Local coral reef ecosystems across the western Puerto Rico shelf still represent a critical natural

resource with invaluable socio-economic significance. But existing impacts from LBSP can not

be sustained for much time and need to be thoroughly addressed through an appropriate, highly-

participatory management plan to reduce LBSP threats. But such a plan needs to be coupled to a

scientific question-driven and properly designed coastal water quality program and a long-term

ecological monitoring program to quantify the long-term impacts of management strategies. This

becomes more critical considering that chronic eutrophication has been shown to increase the

prevalence of coral diseases and bleaching (Vega-Thurber et al., 2013). Considering current and

forecasted climate change-related impacts this will become an even major concern in the near

future. Otherwise, we might be at risk of losing one of the most important coral reef ecosystems

of the entire northeastern Caribbean region.

VI. ACKNOWLEDGMENTS

This study was possible thanks to funding provided by Protectores de Cuencas, Inc. and Ridge to

Reefs, Inc. to Sociedad Ambiente Marino under a Coral Reef Conservation grant from the

National Fish and Wildlife Foundation (NFWF). Partial support was also provided by the

National Science Foundation (HRD #0734826) through the Center for Applied Tropical Ecology

and Conservation (CATEC), and the University of Puerto Rico’s Central Administration to E.A.

Hernández-Delgado. Our appreciation for the logistical field support provided by the crew of

M/V Tourmarine, and Capt. Elick Hernández. This publication is a contribution from CATEC’s

Coral Reef Research Group and SAM’s collaborative Coral Reefs Conservation and

Rehabilitation Project.


131

VII. LITERATURE CITED

Anderson, M.J., Gorley, R.N., and Clarke, K.R. (2008) PERMANOVA + for PRIMER: Guide to

Software and Statistical Methods. PRIMER-E: Plymouth, U.K.

Bejarano-Rodríguez, I., Appeldoorn, R. (2013) Seawater turbidity and fish communities on coral

reefs of Puerto Rico. Marine Ecology Progress Series, 474, 217-226.

Bonkosky, M., Hernández-Delgado, E.A., Sandoz, B., Robledo, I.E., Norat-Ramírez, J., and

Mattei, H. (2009) Detection of Spatial Fluctuations of Non-Point Source Fecal Pollution in

Coral Reef Surrounding Waters in Southwestern Puerto Rico Using PCR-Based Assays.

Marine Pollution Bulletin, 58, 45-54.

Clarke, K.R., and Warwick, R.M. (2001) Change in Marine Communities: An Approach to

Statistical Analysis and Interpretation. 2nd

Ed. PRIMER-E, Ltd., Plymouth Marine

Laboratory, UK.

Cloern, J. (2001) Our Evolving Conceptual Model of the Coastal Eutrophication

Problem. Marine Ecology Progress Series, 210, 223--253.

García-Urueña, R.P. (2004) Dinámica de Crecimiento de Tres Especies de Coral en Relación a

las Propiedades Ópticas del Agua. Ph.D. Dissertation, Department of Marine Sciences,

University of Puerto Rico, Mayagüez, PR, 138 pp.

Goenaga, C. (1988) The Distribution and Growth of Montastrea annularis (Ellis and Solander)

in Puerto Rican platform reefs. Ph.D. Dissertation, Department of Marine Sciences,

University of Puerto Rico, Mayagüez, PR, 215 pp.

Goenaga, C., and Cintrón, G. (1979) Inventory of the Puerto Rican Coral Reefs. Departamento

de Recursos Naturales & National Oceanic and Atmospheric Administration, San Juan, PR,

189 pp.

Goenaga, C., and Canals, M. (1990) Island-Wide Coral Bleaching in Puerto Rico: 1990.

Caribbean Journal of Science, 26, 171-175.

Goenaga, C., and Boulon, R. (1992) The State of Puerto Rican Corals: An Aid to Managers.

Technical Report, Caribbean Fisheries Management Council, San Juan, PR, 66 pp.


132

Hernández-Delgado, E.A. (2000) Effects of Anthropogenic Stress Gradients in the Structure of

Coral Reef Epibenthic and Fish Communities. Ph.D. Dissertation, Department of Biology,

University of Puerto Rico, San Juan, P.R. 330 pp.

Hernández-Delgado, E.A. 2005. Historia natural, caracterización, distribución y estado actual de

los arrecifes de coral Puerto Rico. 281-356. In, R.L. Joglar (Ed.), Biodiversidad de Puerto

Rico: Vertebrados Terrestres y Ecosistemas. Serie Historia Natural. Editorial Instituto de

Cultura Puertorriqueña, San Juan, PR. 563 pp.

Hernández-Delgado, E.A., Sandoz, B., Bonkosky, M., Mattei, H., and Norat, J. (2010) Impacts

of Non-Point Source Sewage Pollution in Elkhorn Coral, Acropora palmata (Lamarck),

Assemblages of the Southwestern Puerto Rico Shelf. Proceedings of the 11th

International

Coral Reefs Symposium, 747-751.

Hernández-Delgado, E.A., and Sandoz-Vera, B. (2011) Impactos Antropogénicos en los

Arrecifes de Coral, 62-72. In, J. Seguinot-Barbosa (Ed.), Islas en Extinción: Impactos

Ambientales en las Islas de Puerto Rico. Ediciones SM, Cataño, PR 255 pp.

Hernández-Delgado, E.A., Hutchinson-Delgado, Y.M., Laureano, R., Hernández-Pacheco, R.,

Ruiz-Maldonado, T.M., Oms, J., and Díaz, P.L. (2011a) Sediment Stress, Water Turbidity

and Sewage Impacts on Threatened Elkhorn Coral (Acropora palmata) Stands at Vega Baja,

Puerto Rico. Proceedings of the Gulf and Caribbean Fisheries Institute, 63, 83-92.

Hernández-Delgado, E.A., Alvarado, A., Laureano, R., Flynn, K., and Griffin, S. (2011b)

Seawall Construction Activities Cause a Localized Mass Mortality of Threatened Elkhorn

Coral (Acropora palmata) at Vega Baja, Puerto Rico. Proceedings of the Gulf and Caribbean

Fisheries Institute, 63, 511.

Hernández-Delgado, E.A., Méndez-Lázaro, P.A., Medina-Muñiz, J.L., Nieves-Santiago, A.,

Cabrera-Molina, J., and Rodríguez-Díaz, I.M. (2012) Chronic Impacts of Sewage and

Turbidity on Coral Reefs Within the Tres Palmas Marine Reserve, Rincón, Puerto Rico: I.

Threatened Elkhorn coral (Acropora palmata) Populations. Final report submitted to

Environmental Science Graduate Program, Department of Public Health, University of

Puerto Rico – Medical Sciences Campus, and to the Coral Reef Conservation Program,

Department of Natural and Environmental Resources, San Juan, Puerto Rico, 70 pp.


133

Hughes, T.P. (1994) Catastrophes, Phase Shifts, and Large-Scale Degradation of a Caribbean

Coral Reef. Science, 265, 1547-1550.

Kaczmarsky, L., Draud, M., and Williams, E. (2005) Is There a Relationship Between Proximity

to Sewage Effluent and the Prevalence of Coral Disease. Caribbean Journal of Science, 41,

124--137.

Lapointe, B.E., and Clark, M.E. (1992) Nutrient Inputs from the Watershed and Coastal

Eutrophication in the Florida Keys. Estuaries, 15, 465-476.

Lapointe, B.E., and Matzie, W.R. (1996) Effects of Stormwater Discharges on Eutrophication

Processes in Nearshore Waters of the Florida Keys. Estuaries, 19, 422-435.

Loya, Y. (1976) Effects of Water Turbidity and Sedimentation on the Community Structure of

Puerto Rican Corals. Bulletin of Marine Science, 26, 450-466.

Meesters, E.H., Bak, R.P.M., Westmacott, S., Ridgley, M., Dollar, S. (1998) A Fuzzy Logic

Model to Predict Coral Reef Development Under Nutrient and Sediment Stress.

Conservation Biology, 12, 957-965.

Méndez-Lázaro, P., Hernández Delgado, E.A., Norat-Ramírez, J., Cordero-Rivera, L., and

Santiago-Nieves, A. (2012) Técnicas de SIG Aplicadas al Estudio de Zonas Costaneras y

Cuencas Hidrográficas en la Isla de Puerto Rico. Revista Internacional de Ciencia y

Tecnología de Información Geogfráfica, Geo-Focus, 12, 71-92.

Norat-Ramírez, J., Méndez-Lázaro, P., Hernández-Delgado, E.A., and Cordero-Rivera, L. (2013)

El Impacto de Aguas Usadas de Fuentes Dispersas en el Litoral Costero de la Reserva

Marina Tres Palmas (Rincón-Puerto Rico). Revista de la Asociación Venezolana de

Ingeniería Sanitaria y Ambiental, 7, 27-30.

Otero, E. (2009) Spatial and Temporal Patterns of Water Quality Indicators in Reef Systems of

Southwestern Puerto Rico. Caribbean Journal of Science, 45, 168-180.

Pastorok, R., and Bilyard, G. 1985. Effects of Sewage Pollution on Coral-Reef Communities.

Marine Ecology Progress Series, 21,175--189.

Schärer, M., Nemeth, M., Valdivia, A., Miller, M., Williams, D., and Diez, C. (2010) Elkhorn

Coral Distribution and Condition Throughout the Puerto Rican Archipelago. Proceedings of

the 11th

International Coral Reef Symposium, 815-819.


134

Sekar, R., Kaczmarsky, L. and Richardson, L. (2008). Microbial Community Composition of

Black Band Disease on the Coral Host Siderastrea siderea from Three Regions of the Wider

Caribbean. Marine Ecology Progress Series, 362, 85-98.

Szmant, A. (2002) Nutrient Enrichment on Coral Reefs: Is It a Major Cause of Coral Reef

Decline?. Estuaries, 25, 743-766.

Tomascik, T., and Sander, F. (1985) Effects of Eutrophication on Reef-Building Corals. I.

Growth Rate of the Reef-Building Coral Montastrea annularis. Marine Biology, 87, 143-155.

Tomascik, T., and Sander, F. (1987) Effects of Eutrophication on Reef-Building Corals. III.

Reproduction of the Reef-Building Coral Porites porites. Marine Biology, 94, 77-84.

Tomascik, T. (1991) Settlement Patterns of Caribbean Scleractinian Corals on Artificial

Substrata Along a Eutrophication Gradient, Barbados, West Indies. Marine Ecology Progress

Series, 77, 261-269.

Torres, J.L., and Morelock, J. (2002) Effect of Terrigenous Sediment Influx on Coral Cover and

Linear Extension Rates of Three Caribbean Massive Coral Species. Caribbean Journal of

Science, 38, 222-229.

Vega-Thurber, R., Burkepile, D., Fuchs, C., Shantz, A., McMinds, R., and Zaneveld, J. (2014)

Chronic Nutrient Enrichment Increases Prevalence and Severity of Coral Disease and

Bleaching. Global Change Biology, 20, 544--554.

Walker, D., and Ormond, R. (1982) Coral Death from Sewage and Phosphate Pollution at Aqaba,

Red Sea. Marine Pollution Bulletin, 13, 21-25.

Weil, E., Hernández-Delgado, E., Bruckner, A., Ortiz, A., Nemeth, M., and Ruiz, H. (2002)

Distribution and Status of Acroporid Coral (Scleractinia) Populations in Puerto Rico, 71-98.

In, Bruckner, A.W. (ed.), Proceedings of the Caribbean Acropora Workshop: Potential

Application of the U.S. Endangered Species Act as a Conservation Strategy. NOAA Tech.

Memorandum NMFS-OPR-24, Silver Spring, MD, 199 pp.

Widespread Impacts of Land Based Source …...Widespread Impacts of Land-Based Source Pollution on...

Documents

Transcript of Widespread Impacts of Land Based Source …...Widespread Impacts of Land-Based Source Pollution on...